+2

languages/ziglang/0.15/README.md

reviewed

··· 25 25 - [modules](./modules.md) - file + directory pattern (foo.zig + foo/) 26 26 - [structs](./structs.md) - copy semantics, internal storage for strings 27 27 - [testing](./testing.md) - zig test vs build test, arena for leaky apis 28 28 + - [hashmap](./hashmap.md) - StringHashMap, O(1) block index, managed vs unmanaged 29 29 + - [format](./format.md) - {f} required for format methods, {t} for tag names, signature change

+42

languages/ziglang/0.15/binary.md

reviewed

··· 43 43 try encode(alloc, list.writer(alloc), value); 44 44 ``` 45 45 46 46 + **note**: `std.io.fixedBufferStream` is deprecated in 0.15 — the stdlib says to use `std.Io.Writer.fixed` / `std.Io.Reader.fixed` instead. the old API still compiles (zat uses it in 3 files) but new code should prefer the non-deprecated form. the `anytype` writer pattern itself is fine either way — the encoder doesn't care which writer type backs it. 47 47 + 46 48 see: [zat/cbor.zig](https://tangled.sh/@zzstoatzz.io/zat/tree/main/src/internal/cbor.zig) 47 49 48 50 ## encodeAlloc convenience ··· 133 135 this means the handler's `onEvent` must not hold references to event data past the call. if it needs to, it must copy into its own allocator. 134 136 135 137 see: [zat/firehose.zig](https://tangled.sh/@zzstoatzz.io/zat/tree/main/src/internal/firehose.zig), [zat/jetstream.zig](https://tangled.sh/@zzstoatzz.io/zat/tree/main/src/internal/jetstream.zig) 138 138 + 139 139 + ## specialized decoders 140 140 + 141 141 + when generic decoding is too expensive, write a purpose-built parser for a known schema. the generic path builds `Value` unions, `MapEntry` arrays, and handles every CBOR type. if you know the exact shape, skip all that. 142 142 + 143 143 + example: MST nodes are always `map(2) { "e": array[entries...], "l": CID|null }`. instead of `cbor.decodeAll()` → extract fields from Value unions, parse the CBOR bytes directly: 144 144 + 145 145 + ```zig 146 146 + pub fn decodeMstNode(allocator: Allocator, data: []const u8) MstDecodeError!MstNodeData { 147 147 + // expect map(2), key "e", array(n) — known byte sequence 148 148 + // parse entries inline, zero-copy slicing into input buffer 149 149 + // only allocation: the entries array itself 150 150 + } 151 151 + 152 152 + pub const MstNodeData = struct { 153 153 + left: ?[]const u8, // raw CID bytes (borrowed from input) 154 154 + entries: []MstEntryData, // heap-allocated array 155 155 + }; 156 156 + 157 157 + pub const MstEntryData = struct { 158 158 + prefix_len: usize, 159 159 + key_suffix: []const u8, // borrowed from input 160 160 + value_cid: []const u8, // borrowed from input 161 161 + tree: ?[]const u8, // borrowed from input 162 162 + }; 163 163 + ``` 164 164 + 165 165 + the result: MST walk went from 45.5ms (generic decode per node) to 39.3ms (specialized decode) on 243k blocks. the bigger win was avoiding the full tree rebuild (218ms → 39ms total) by verifying structure during the walk. 166 166 + 167 167 + when to use this pattern: 168 168 + - you decode the same schema thousands of times (MST nodes, CBOR blocks) 169 169 + - the schema is stable and well-known 170 170 + - profiling shows decode time dominates 171 171 + 172 172 + when NOT to use it: 173 173 + - the schema varies or is user-defined 174 174 + - you only decode a handful of times 175 175 + - generic decode is fast enough 176 176 + 177 177 + see: [zat/mst.zig decodeMstNode](https://tangled.sh/@zzstoatzz.io/zat/tree/main/src/internal/repo/mst.zig) 136 178 137 179 ## deterministic encoding 138 180

+23

languages/ziglang/0.15/build.md

reviewed

··· 41 41 ``` 42 42 43 43 without this line, `zig build` runs successfully but produces no output. easy to miss. 44 44 + 45 45 + ## x86 backend (default in 0.15) 46 46 + 47 47 + the self-hosted x86 backend is now the **default** for debug builds. roughly 5x faster compilation than LLVM for most code. threaded codegen adds up to 50% more on top of that. 48 48 + 49 49 + if you hit codegen bugs, fall back to LLVM: 50 50 + 51 51 + ```bash 52 52 + zig build -fllvm # use LLVM backend 53 53 + zig build -Doptimize=ReleaseFast # release modes always use LLVM 54 54 + ``` 55 55 + 56 56 + the x86 backend passes more behavior tests than LLVM (1984 vs 1977) but generates slower machine code. debug builds prioritize compilation speed; release builds prioritize runtime performance. 57 57 + 58 58 + ## test-obj 59 59 + 60 60 + 0.15 added `zig test-obj` for compiling tests to object files instead of running them. useful when linking test code into external harnesses: 61 61 + 62 62 + ```bash 63 63 + zig test-obj src/foo.zig # produces .o file 64 64 + ``` 65 65 + 66 66 + in build.zig: `addTest(.{ ... }).emit_object = true;`

+23

languages/ziglang/0.15/crypto.md

reviewed

··· 43 43 44 44 see: [zat/jwt.zig](https://tangled.sh/@zzstoatzz.io/zat/tree/main/src/internal/jwt.zig) 45 45 46 46 + ## signing 47 47 + 48 48 + ```zig 49 49 + const Scheme = std.crypto.sign.ecdsa.EcdsaSecp256k1Sha256; 50 50 + 51 51 + // secret key from raw 32 bytes 52 52 + const secret_key = try Scheme.SecretKey.fromBytes(key_bytes[0..32].*); 53 53 + 54 54 + // sign with RFC 6979 deterministic nonce (null = no additional randomness) 55 55 + const sig = secret_key.sign(message, null); 56 56 + 57 57 + // low-S normalization (required by AT Protocol / Bitcoin) 58 58 + const Curve = std.crypto.ecc.Secp256k1; 59 59 + const s_order = Curve.scalar.neg(sig.s, .big); 60 60 + // if s > half_order, replace with order - s 61 61 + ``` 62 62 + 63 63 + the `null` second arg means RFC 6979 deterministic nonce — no randomness source needed, signature is reproducible from the same key+message. 64 64 + 65 65 + low-S normalization ensures `s <= half_order`. some protocols (AT Protocol, Bitcoin) reject high-S signatures. check by comparing the s component against the curve half-order. 66 66 + 67 67 + see: [zat/jwt.zig](https://tangled.sh/@zzstoatzz.io/zat/tree/main/src/internal/crypto/jwt.zig), [k256 tests](https://tangled.sh/@zzstoatzz.io/k256) 68 68 + 46 69 ## sha-256 hashing 47 70 48 71 for content addressing (CIDs, integrity checks). hash data into a fixed-size digest:

+5 -4

languages/ziglang/0.15/database.md

reviewed

··· 251 251 252 252 ```zig 253 253 fn rewritePlaceholders(alloc: Allocator, sql: []const u8) ![]const u8 { 254 254 - var result = std.ArrayList(u8).init(alloc); 254 254 + var result: std.ArrayList(u8) = .empty; 255 255 + errdefer result.deinit(alloc); 255 256 var param_num: usize = 1; 256 257 257 258 for (sql) |c| { 258 259 if (c == '?') { 259 259 - try result.writer().print("${d}", .{param_num}); 260 260 + try result.print(alloc, "${d}", .{param_num}); 260 261 param_num += 1; 261 262 } else { 262 262 - try result.append(c); 263 263 + try result.append(alloc, c); 263 264 } 264 265 } 265 265 - return result.toOwnedSlice(); 266 266 + return result.toOwnedSlice(alloc); 266 267 } 267 268 ``` 268 269

+61

languages/ziglang/0.15/format.md

reviewed

··· 1 1 + # format strings 2 2 + 3 3 + 0.15 changed how `std.fmt` works. the big ones: `{f}` is now required for types with format methods, and format method signatures changed. 4 4 + 5 5 + ## {f} required for format methods 6 6 + 7 7 + in 0.14, `{}` would implicitly call a type's `format` method if it had one. in 0.15, you must use `{f}` explicitly: 8 8 + 9 9 + ```zig 10 10 + const MyType = struct { 11 11 + value: u32, 12 12 + 13 13 + pub fn format(self: @This(), writer: *std.Io.Writer) std.Io.Writer.Error!void { 14 14 + try writer.print("MyType({d})", .{self.value}); 15 15 + } 16 16 + }; 17 17 + 18 18 + const x = MyType{ .value = 42 }; 19 19 + 20 20 + // 0.14: std.debug.print("{}", .{x}); // called format implicitly 21 21 + // 0.15: std.debug.print("{f}", .{x}); // must use {f} 22 22 + // std.debug.print("{any}", .{x}); // skips format method 23 23 + ``` 24 24 + 25 25 + why: prevents silent behavior changes when format methods are added or removed from types. 26 26 + 27 27 + ## format method signature change 28 28 + 29 29 + ```zig 30 30 + // 0.14 signature 31 31 + pub fn format( 32 32 + self: @This(), 33 33 + comptime fmt_string: []const u8, 34 34 + options: std.fmt.FormatOptions, 35 35 + writer: anytype, 36 36 + ) !void 37 37 + 38 38 + // 0.15 signature 39 39 + pub fn format( 40 40 + self: @This(), 41 41 + writer: *std.Io.Writer, 42 42 + ) std.Io.Writer.Error!void 43 43 + ``` 44 44 + 45 45 + no more format string or options parameters. if you need different formatting modes, use separate named methods or return a wrapper struct. 46 46 + 47 47 + ## new specifiers 48 48 + 49 49 + | specifier | meaning | example | 50 50 + |-----------|---------|---------| 51 51 + | `{f}` | call type's format method | `std.debug.print("{f}", .{my_obj})` | 52 52 + | `{t}` | `@tagName()` / `@errorName()` | `std.debug.print("{t}", .{my_enum})` | 53 53 + | `{b64}` | base64 encoding | `std.debug.print("{b64}", .{bytes})` | 54 54 + 55 55 + `{t}` is particularly useful — replaces `@tagName(enum_val)` in format strings. 56 56 + 57 57 + ## what still works 58 58 + 59 59 + `{s}`, `{d}`, `{x}`, `{any}`, `{?}`, `{}` for basic types are unchanged. our zat code uses these exclusively and doesn't need migration. 60 60 + 61 61 + sources: [0.15 release notes](https://ziglang.org/download/0.15.1/release-notes.html)

+79

languages/ziglang/0.15/hashmap.md

reviewed

··· 1 1 + # hashmap 2 2 + 3 3 + zig's hash maps for key-value lookup. two flavors parallel the arraylist split: `StringHashMap` stores the allocator, `StringHashMapUnmanaged` passes it to each method. 4 4 + 5 5 + ## managed vs unmanaged 6 6 + 7 7 + ```zig 8 8 + // managed — stores allocator, slightly more convenient 9 9 + var map = std.StringHashMap([]const u8).init(allocator); 10 10 + defer map.deinit(); 11 11 + 12 12 + try map.put("key", "value"); 13 13 + 14 14 + // unmanaged — pass allocator to each method (like ArrayList) 15 15 + var map: std.StringHashMapUnmanaged([]const u8) = .empty; 16 16 + defer map.deinit(allocator); 17 17 + 18 18 + try map.put(allocator, "key", "value"); 19 19 + ``` 20 20 + 21 21 + unmanaged is the better default for the same reasons as ArrayList — no stored allocator, statically initializable. use managed when passing as function parameter (avoids threading the allocator alongside it). 22 22 + 23 23 + ## O(1) block index 24 24 + 25 25 + the pattern that fixed zat's 79-second MST walk: build a lookup table during CAR parse, query it during tree traversal. 26 26 + 27 27 + ```zig 28 28 + pub const Car = struct { 29 29 + blocks: []const Block, 30 30 + block_index: std.StringHashMapUnmanaged([]const u8) = .empty, 31 31 + }; 32 32 + 33 33 + // during parse: populate index alongside the block array 34 34 + var block_index: std.StringHashMapUnmanaged([]const u8) = .empty; 35 35 + while (pos < data.len) { 36 36 + // ... parse CID and content ... 37 37 + try blocks.append(allocator, .{ .cid_raw = cid_bytes, .data = content }); 38 38 + try block_index.put(allocator, cid_bytes, content); 39 39 + } 40 40 + 41 41 + // during walk: O(1) lookup by CID 42 42 + const data = block_index.get(cid_bytes) orelse return error.BlockNotFound; 43 43 + ``` 44 44 + 45 45 + the key insight: CID bytes are already unique identifiers and the block data they point to is owned by the CAR buffer (arena-allocated). no copies needed — both key and value are slices into existing memory. 46 46 + 47 47 + before this: linear scan through 243k blocks, called once per MST node (~50k nodes). ~12 billion comparisons. after: hash lookup. 48 48 + 49 49 + see: [zat/car.zig](https://tangled.sh/@zzstoatzz.io/zat/tree/main/src/internal/repo/car.zig) 50 50 + 51 51 + ## key ownership 52 52 + 53 53 + from the stdlib: "key memory is managed by the caller. keys and values will not automatically be freed." 54 54 + 55 55 + this means: 56 56 + - keys must outlive the map 57 57 + - if keys are heap-allocated, you free them — the map won't 58 58 + - for string keys pointing into a long-lived buffer (like parsed data), this is free — the buffer already owns the memory 59 59 + 60 60 + ## HashMap vs ArrayHashMap 61 61 + 62 62 + `HashMap` (and `StringHashMap`): hash-table backed. O(1) lookup, unordered iteration. 63 63 + 64 64 + `ArrayHashMap` (and `StringArrayHashMap`): array-backed with hash index. O(1) lookup, **insertion-order** iteration, slightly more memory. 65 65 + 66 66 + use HashMap when you only care about lookup. use ArrayHashMap when iteration order matters (e.g., rendering parameters in a URL). 67 67 + 68 68 + ## generic keys with AutoHashMap 69 69 + 70 70 + for non-string keys, `AutoHashMap` generates hash/eql from the key type: 71 71 + 72 72 + ```zig 73 73 + var map = std.AutoHashMap(u64, MyStruct).init(allocator); 74 74 + defer map.deinit(); 75 75 + 76 76 + try map.put(42, my_value); 77 77 + ``` 78 78 + 79 79 + works for integers, enums, and types with `hash` and `eql` methods.

+2

languages/ziglang/README.md

reviewed

··· 20 20 | [pollz](https://tangled.sh/@zzstoatzz.io/pollz) | bluesky polls (zqlite + transactions) | 21 21 | [zql](https://tangled.sh/@zzstoatzz.io/zql) | comptime sql parsing | 22 22 | [zat](https://tangled.sh/@zzstoatzz.io/zat) | atproto primitives (jwt, crypto, CBOR, firehose) | 23 23 + | [k256](https://tangled.sh/@zzstoatzz.io/k256) | optimized secp256k1 ECDSA (5×52-bit field, GLV endomorphism) | 24 24 + | [atproto-bench](https://tangled.sh/@zzstoatzz.io/atproto-bench) | three-way AT Protocol benchmarks (zig vs Go vs Rust) | 23 25 | [logfire-zig](https://tangled.sh/@zzstoatzz.io/logfire-zig) | OTLP observability client | 24 26 | [prefect-zig](https://tangled.sh/@zzstoatzz.io/prefect-zig) | prefect orchestration server | 25 27 | [ghostty](https://github.com/ghostty-org/ghostty) | terminal emulator (build system) |

+1

protocols/atproto/README.md

reviewed

··· 67 67 - [firehose](./firehose.md) - event streaming, jetstream 68 68 - [auth](./auth.md) - OAuth, scopes, permission sets 69 69 - [labels](./labels.md) - moderation, signed assertions 70 70 + - [sync-verification](./sync-verification.md) - inductive proof chains, MST inversion, sync 1.1 70 71 71 72 ## sources 72 73

+148

protocols/atproto/dev-chats/02-22-26-writeup.md

reviewed

··· 1 1 + ATProtoDevChat part IV (02/22/2026) 2 2 + 3 3 + i am human and i err! this was rendered from memory and discord chat, pls comment/DM me if you see mistakes or misrepresentations 4 4 + 5 5 + ## Topics Covered 6 6 + 7 7 + ### Labelers: Skyware, Ozone & Osprey 8 8 + 9 9 + @byarielm opened with "what's the difference between Skyware and Ozone?" — a question many newcomers hit when approaching the moderation stack. 10 10 + 11 11 + The short answer: Skyware is a minimal labeler implementation — it issues labels and sends them all to the client (no incremental backfill). Ozone is a full multi-user moderation interface for human review. And Osprey is the newest piece: an automation and data analysis engine for pattern detection at scale. 12 12 + 13 13 + @thisismissem.social explained that Osprey "receives the events from a source, e.g., the firehose, and then based on the rules you've stored (a DSL) it can then apply those to data as it comes in — it's very similar in that way to SQRL. Osprey also stores all the events in a big database, which then allows you to write new rules and investigate patterns." 14 14 + 15 15 + @byarielm summarized: "Skyware is kinda the base code you need for a labeler, Ozone provides UI and some extra db/moderation stuff, and Osprey is more pattern/rule formatting for labellers?" 16 16 + 17 17 + Key resources: 18 18 + - [Osprey for ATProto](https://github.com/haileyok/osprey-for-atproto) 19 19 + 20 20 + ### Signup UX & prompt=create 21 21 + 22 22 + @flo.bit raised the signup UX question: should apps let users enter email and password on the app's own page (easier to understand, but worse for password managers — this is what Bluesky does), or redirect them to a PDS with `prompt=create`? 23 23 + 24 24 + The problem is that `prompt=create` isn't standard and not all PDSs support it — single-user PDSs and those with registration disabled can't handle it. Updates to both the PDS and the OAuth flow itself are needed. There is reportedly energy on the Bluesky team around this. 25 25 + 26 26 + @thisismissem.social shared her OAuth working group proposal, which addresses some of these authorization management concerns. 27 27 + 28 28 + Key resources: 29 29 + - [OAuth Authorization Management URI (draft)](https://github.com/ThisIsMissEm/draft-oauth-authorization-management-uri) 30 30 + - [Bluesky Cookbook](https://github.com/bluesky-social/cookbook/) 31 31 + 32 32 + ### Device Code Grant Flow 33 33 + 34 34 + Brief discussion of the device code grant flow pattern: start the flow on one device, poll for completion, and get the token back. The open question is how DPoP works with device code grant flow (still in draft), and what happens for devices without a display. 35 35 + 36 36 + @same.supply shared an example implementation from the atcute project. 37 37 + 38 38 + Key resources: 39 39 + - [atcute node-client-public-example](https://github.com/mary-ext/atcute/tree/trunk/packages/oauth/node-client-public-example) 40 40 + 41 41 + ### Explaining Credible Exit to Non-Technical Users 42 42 + 43 43 + This thread explored the perennial challenge: how do you explain PLC, rotation keys, and credible exit to people who aren't developers? 44 44 + 45 45 + @sharpie offered a framing that landed well: "PLC is a cloud thing. It stores the public part of your rotation key, allowing you to move your data home if your house burns down." The group riffed on analogies — @sorryforpartyrock.ing compared rotation keys to 2FA backup keys, @byarielm suggested "it's like lock-and-key matching but more like fingerprints that are actually unique," and @psingletary.com offered the image of BFF heart necklaces — one half yours, one half somewhere else. 46 46 + 47 47 + @thisismissem.social took a pragmatic stance: "I'd just say you have a credential that you can use to manage your identity in the cloud, and when you create that credential, you register it with your identity in the cloud." 48 48 + 49 49 + The challenge, as @sharpie put it: "getting people to tune in long enough to even understand what the PLC is — is very hard. I can hear people's eyes glazing over." @geesawra noted simply: "asymmetric crypto is hard." 50 50 + 51 51 + Discussion also touched on whether rotation keys should be surfaced in Bluesky's settings under an "Advanced" menu. @geesawra argued for it ("otherwise I feel they'll always be reserved for nerds like me"), while @same.supply felt they should live on your PDS host's website, not in any particular app. @dane agreed it shouldn't be a separate thing from the PDS. The secure enclave approach — storing keys in device hardware — was mentioned as the ideal end state. 52 52 + 53 53 + @phil shared a post on adversarial PLC directory migration, rounding out the conversation with the question of what happens if the centralized PLC directory itself becomes hostile. 54 54 + 55 55 + Key resources: 56 56 + - [Adversarial PLC directory migration](https://updates.microcosm.blue/3lz7nwvh4zc2u) 57 57 + - [Sharpie's PDS migration vlog](https://bsky.app/profile/sharpiepls.com/post/3lw3hcuojck2u) 58 58 + 59 59 + ### Everything Account 60 60 + 61 61 + Brief but fun thread: @geesawra floated the idea of putting dotfiles on your PDS — "ultimate nerd snipe." @quillmatiq.com named the vision: "everything account." @evan.jarrett.net pointed out the real draw would be if the data could be encrypted/private, and @same.supply noted that dotfiles are usually already public on git anyway. The encrypted/private data angle keeps pulling every conversation toward it. 62 62 + 63 63 + ### Relays & collectiondir 64 64 + 65 65 + Quick pointers to relay infrastructure resources. @thisismissem.social shared the AT Stack guide. @dane linked collectiondir from the Indigo repo — a tool for browsing collections on the relay. @verdverm.com mentioned their atmunge project, which "produces a table with every account and collections — it's very easy to build that index, one request per account." 66 66 + 67 67 + Key resources: 68 68 + - [The AT Stack (guide)](https://atproto.com/guides/the-at-stack) 69 69 + - [collectiondir (Indigo)](https://github.com/bluesky-social/indigo/tree/main/cmd/collectiondir) 70 70 + 71 71 + ### Lexicons 72 72 + 73 73 + A pragmatic conversation about when you need a lexicon and when you don't. @flo.bit acknowledged that Blento "still doesn't have a lexicon" and @psingletary.com formalized the vibe: "Not Everything Needs A Lexicon." @quillmatiq.com agreed — "we can also start sans lexicon and move when we're ready, doesn't need to be a blocker." 74 74 + 75 75 + That said, when you do use lexicons, the tooling needs work. @geesawra called the deprecated field support "vital." @taurean.bryant.land pointed to Lexicon Garden as the best current resource for browsing existing schemas. 76 76 + 77 77 + Key resources: 78 78 + - [Lexicon Garden](https://lexicon.garden/) 79 79 + 80 80 + ### Documentation & Contributions 81 81 + 82 82 + The "circle of life" problem: @evan.jarrett.net "tried to write documentation and then it got outdated like 3 commits later." This resonated broadly. 83 83 + 84 84 + The aspirational model: @thoth pushed for rustdoc-style generation from source annotations — "I didn't think it was possible for docs generated from source annotations to be good (I mean, have you seen doxygen?) until I saw rustdoc." @geesawra added godoc to the list. 85 85 + 86 86 + @byarielm and @chaosgreml.in (Bryan Guffey) found each other over wanting "for babies" / ELI5 docs and volunteered to collaborate. @n8 framed the need as three tiers: "examples/concepts/api-ref, I think we need all 3, some will be app level." @same.supply's pattern from JS libraries: "Getting Started/Quick Start + Important Concepts + API Type specification." @geesawra pointed to Go's docs as a model: "examples/docs/howtos plus reference in a single page." 87 87 + 88 88 + On the contribution side, @thisismissem.social pointed to the atproto-website GitHub issues and @n8 noted the team "were pretty responsive on GH, got a little typo thing merged in a couple hours." @byarielm's workflow challenge: "it's hard to read and then pause to write an issue." @thisismissem.social suggested annotating docs with margin.at as a lower-friction way to leave notes while reading. 89 89 + 90 90 + Key resources: 91 91 + - [atproto-website issues](https://github.com/bluesky-social/atproto-website/issues) 92 92 + 93 93 + ### Following Feeds 94 94 + 95 95 + The following feed is "a stupidly complex data problem at scale" — and according to a post from @why.bsky.team, it represents roughly half of Bluesky's production workload. The basic flow: a post lands, a background task injects it into every follower's feed (a sorted set). Straightforward in concept, crushing at scale. 96 96 + 97 97 + @evan.jarrett.net clarified this is specifically about "the non-algorithmic one, because you have to show every post for every follower that everyone has." Mastodon's approach: generate the following feed on demand and suspend it after inactivity. Bluesky uses ScyllaDB and does not suspend feed generation for inactive users — which contributes to the workload. The Postgres-to-Scylla migration reportedly went poorly and touched everything. 98 98 + 99 99 + Large accounts may be lossy — if you follow enough people, you might not see everything. @byarielm linked Jaz's post on "imperfection" as relevant background. 100 100 + 101 101 + For app developers wanting to build their own following feed without all this infrastructure, there's a cheat: take {bsky follows} minus {those without your app's *.xyz records} — essentially filtering your following list to only people who use your app. 102 102 + 103 103 + Key resources: 104 104 + - [Why on following feed workload](https://witchsky.app/profile/did:plc:vpkhqolt662uhesyj6nxm7ys/post/3m2wsfeab322b) 105 105 + - [Jaz — Imperfection](https://jazco.dev/2025/02/19/imperfection/) 106 106 + 107 107 + ### Blocks Don't Remove Follow Records 108 108 + 109 109 + An important gotcha surfaced in the following feed discussion: blocking someone on Bluesky does not delete the follow record from your repo. @geesawra noted that "the good old 'block, unblock to force unfollow' trick doesn't work around here." 110 110 + 111 111 + @verdverm.com posed the deeper question: "why should block activity force some unfollow? Or generalized, why does writing a record in one repo force a write in another repo?" — highlighting a fundamental design tension in how repos interact. 112 112 + 113 113 + @taurean.bryant.land offered a practical angle: using the followgraph as a follow recommendation system to build a new social graph, rather than trying to force cross-repo writes. 114 114 + 115 115 + ### Community Showcase 116 116 + 117 117 + **arabica** — @pdewey.com demoed their coffee journaling app built on ATProto (alpha). @tynanpurdy.com couldn't help wondering if coffee tasting notes and fragrance review notes could share a lexicon. 118 118 + 119 119 + **drydown** — @taurean.bryant.land shared their fragrance review app built on ATProto. @sharpie expressed interest in discussing creator use cases together. 120 120 + 121 121 + **ATmosphereConf** — @psingletary.com ran a poll: 42% virtual, 33% IRL, 17% "what's that?", 8% nah. The conference is in Vancouver, March 26-29, 2026. Several attendees expressed they'd attend virtually due to not feeling comfortable leaving the US right now. 122 122 + 123 123 + **atproto table at local events** — @sharpie mentioned hosting an ATProto table at their local event the following week. 124 124 + 125 125 + **OG/link preview issues** — @michaelthatsit raised issues with Bluesky's handling of Open Graph previews for dynamically generated images. @sri.xyz pointed to the cardyb endpoint that Bluesky uses for link extraction. @sorryforpartyrock.ing shared gizo, their link card generator service. 126 126 + 127 127 + **concord** — @alexwykoff shared that @dollspace-gay built an IRC client that uses ATProto auth. 128 128 + 129 129 + Key resources: 130 130 + - [arabica (alpha)](https://alpha.arabica.social/) 131 131 + - [arabica on Tangled](https://tangled.org/did:plc:chqc2ockzmyvlrasfb66x64a/arabica) 132 132 + - [drydown](https://drydown.social/profile/taurean.bryant.land/house/3mcn34apccq2z) 133 133 + - [ATmosphereConf 2026](https://atmosphereconf.org/) 134 134 + - [cardyb](https://cardyb.bsky.app/v1/extract?url=https://atproto.com) 135 135 + - [gizo](https://github.com/espeon/gizo) 136 136 + - [concord](https://github.com/dollspace-gay/concord) 137 137 + 138 138 + ### Legal Obligations for PDS Operators 139 139 + 140 140 + In the last stretch of the call, @chaosgreml.in (Bryan Guffey) brought up legal compliance for PDS operators — content licensing, record versioning, and the obligations that come with hosting user-generated content. @tynanpurdy.com asked if this would be the impetus for record versioning and content licensing support. 141 141 + 142 142 + @jackvalinsky raised the question of how compliance records should be structured — whether to use site.standard or a dedicated lexicon, given the need for versioning and tombstoning. One approach discussed: using the CID of a record as the rkey when creating it, then linking back on updates, giving you a content-addressed history with a mutable pointer to the current version. 143 143 + 144 144 + The motivating concern, as @jackvalinsky put it: "we need to make it simple and straightforward to be informed because we want some random student to host a project and not be scared because of regulations that they are doing it wrong." 145 145 + 146 146 + ## All links 147 147 + 148 148 + [Osprey for ATProto](https://github.com/haileyok/osprey-for-atproto) ∙ [OAuth Authorization Management URI (draft)](https://github.com/ThisIsMissEm/draft-oauth-authorization-management-uri) ∙ [Bluesky Cookbook](https://github.com/bluesky-social/cookbook/) ∙ [atcute node-client-public-example](https://github.com/mary-ext/atcute/tree/trunk/packages/oauth/node-client-public-example) ∙ [Adversarial PLC directory migration](https://updates.microcosm.blue/3lz7nwvh4zc2u) ∙ [Sharpie's PDS migration vlog](https://bsky.app/profile/sharpiepls.com/post/3lw3hcuojck2u) ∙ [The AT Stack (guide)](https://atproto.com/guides/the-at-stack) ∙ [collectiondir (Indigo)](https://github.com/bluesky-social/indigo/tree/main/cmd/collectiondir) ∙ [Lexicon Garden](https://lexicon.garden/) ∙ [atproto-website issues](https://github.com/bluesky-social/atproto-website/issues) ∙ [Why on following feed workload](https://witchsky.app/profile/did:plc:vpkhqolt662uhesyj6nxm7ys/post/3m2wsfeab322b) ∙ [Jaz — Imperfection](https://jazco.dev/2025/02/19/imperfection/) ∙ [arabica (alpha)](https://alpha.arabica.social/) ∙ [arabica on Tangled](https://tangled.org/did:plc:chqc2ockzmyvlrasfb66x64a/arabica) ∙ [drydown](https://drydown.social/profile/taurean.bryant.land/house/3mcn34apccq2z) ∙ [ATmosphereConf 2026](https://atmosphereconf.org/) ∙ [cardyb](https://cardyb.bsky.app/v1/extract?url=https://atproto.com) ∙ [gizo](https://github.com/espeon/gizo) ∙ [concord](https://github.com/dollspace-gay/concord) ∙ [thisismissem on infra/apps/moderation](https://bsky.app/profile/thisismissem.social/post/3mfg3fc2jwc24)

+1777

protocols/atproto/dev-chats/02-22-26.md

reviewed

··· 1 1 + <notes> 2 2 + # notes 3 3 + 4 4 + asjdlfk 5 5 + 6 6 + ## labelers 7 7 + 8 8 + aggregating 9 9 + 10 10 + skyware: SELECT \* labels -> sends out all labels to client (doesnt do incremental backfill) 11 11 + 12 12 + "what's the difference between skyware and ozone?" 13 13 + 14 14 + - ozone: whole multi-user interface for human moderation 15 15 + - skyware: minimal labeler thing 16 16 + 17 17 + osprey: patterns at scale 18 18 + 19 19 + coop: 20 20 + 21 21 + *** 22 22 + 23 23 + ## UX around prompt create 24 24 + 25 25 + prompt=create not standard, not all PDSs support it (e.g. single user PDS, PDS has user-registration disabled) 26 26 + 27 27 + > updates to the PDS and the oauth flow itself are needed 28 28 + 29 29 + > there is energy on the bsky team around this 30 30 + 31 31 + (oauth working group proposal repo on emelia's repo) 32 32 + 33 33 + ## device code grant flow 34 34 + 35 35 + start flow from one device, poll, token back 36 36 + 37 37 + DPoP with device code grant flow, draft: 38 38 + 39 39 + > devices without a display?? 40 40 + 41 41 + [atcute/packages/oauth/node-client-public-example at trunk · mary-ext/atcute](https://github.com/mary-ext/atcute/tree/trunk/packages/oauth/node-client-public-example) 42 42 + 43 43 + ## relays 44 44 + 45 45 + <https://atproto.com/guides/the-at-stack> 46 46 + 47 47 + ## everything account 48 48 + 49 49 + ## how to explain credible exit 50 50 + 51 51 + why (plc rotation keys in secure enclave), hard to understand 52 52 + 53 53 + sharpie language: "PLC is cloud thing. stores public part of rotation key, allowing you to move your data home if your house burns down" 54 54 + 55 55 + "fingerprint for your account" like 2FA backup keys, which you store yourself 56 56 + 57 57 + ms boba 58 58 + 59 59 + sharpie: partnering with content creators 60 60 + 61 61 + Doctor Popular?? 62 62 + 63 63 + ## lexicons 64 64 + 65 65 + not everything needs to be in lexicons 66 66 + 67 67 + deprecated field 68 68 + 69 69 + docs pointer 70 70 + 71 71 + lexicon garden 72 72 + 73 73 + - what's missing? 74 74 + 75 75 + ## docs contributions 76 76 + 77 77 + circle of life 78 78 + 79 79 + ## following feeds 80 80 + 81 81 + stupidly complex data problem at scale 82 82 + 83 83 + [like half of bskys workload](https://witchsky.app/profile/did:plc:vpkhqolt662uhesyj6nxm7ys/post/3m2wsfeab322b) 84 84 + 85 85 + post > bg task > inject into follower feeds (sorted set) 86 86 + 87 87 + mastodon gens following feed on demand after inactivity 88 88 + 89 89 + postgres -> scylla (went poorly, touched everything) 90 90 + 91 91 + bsky: scylla db, don't suspend creating feeds for inactive users 92 92 + 93 93 + lrg account -> lossy? might not get it? 94 94 + 95 95 + <https://jazco.dev/2025/02/19/imperfection/> 96 96 + 97 97 + > has popfeed done this? 98 98 + 99 99 + cheat: {bsky follows} - {those w/o \*.xyz records} 100 100 + 101 101 + gotcha: blocks DNE removal of follow record 102 102 + 103 103 + </notes> 104 104 + 105 105 + <discord chat> 106 106 + thisismissem.social — 2:00 PM 107 107 + hi ih 108 108 + n8 — 2:00 PM 109 109 + heya 110 110 + Bryan Guffey 111 111 + 112 112 + — 2:00 PM 113 113 + woo woo 114 114 + sri.xyz — 2:01 PM 115 115 + 👋 116 116 + thisismissem.social 117 117 + is now a speaker. — 2:01 PM 118 118 + pdewey.com — 2:01 PM 119 119 + 👋 120 120 + flo.bit — 2:01 PM 121 121 + also have some topic suggestions: 122 122 + 123 123 + long text format: markdown vs lexicon 124 124 + signup: sign up on page vs on pds with prompt=create 125 125 + hi everyone 🙂 126 126 + n8 — 2:02 PM 127 127 + hey flo bit!!!! 128 128 + taking notes: https://leaflet.pub/66b63f56-f335-4d41-8e4a-0eca566cfb22 129 129 + notes 130 130 + asjdlfk 131 131 + notes 132 132 + evan.jarrett.net 133 133 + 134 134 + — 2:07 PM 135 135 + what no gemini to record this conversation and transcribe it for us? 136 136 + byarielm — 2:07 PM 137 137 + What's the difference btwn skyware and ozone? 138 138 + verdverm.com — 2:07 PM 139 139 + fyi, the noise level between speakers is large, n8 quiet, em loud 140 140 + flo.bit — 2:08 PM 141 141 + i started a transcript 142 142 + thisismissem.social — 2:08 PM 143 143 + you can boost individual people's audio 144 144 + byarielm — 2:08 PM 145 145 + you can manually adjust audio for each user! 146 146 + verdverm.com — 2:08 PM 147 147 + been looking, but not seeing 148 148 + byarielm — 2:08 PM 149 149 + click the little three dots after hovering over n8's video 150 150 + pdewey.com — 2:09 PM 151 151 + you can also just right click on a profile picture to see that menu 152 152 + evan.jarrett.net 153 153 + 154 154 + — 2:09 PM 155 155 + or right click 156 156 + flo.bit — 2:09 PM 157 157 + also another topic suggestion: 158 158 + easiest way to build an appview with a following feed 159 159 + verdverm.com — 2:10 PM 160 160 + i s permissioned data on the agenda? 161 161 + n8 — 2:10 PM 162 162 + yea somewhat, i have an item about "private data use cases" inspired by bryan's post 163 163 + verdverm.com — 2:10 PM 164 164 + Bsky is moving over to Osprey? 165 165 + flo.bit — 2:10 PM 166 166 + yeah lets see for what we have the time, all only suggestions if we have nothing else to talk about 😉 167 167 + n8 — 2:11 PM 168 168 + i like to follow whatever comes up instead of being prescriptive about the topics so im game for it 🙂 169 169 + flo.bit — 2:12 PM 170 170 + the signup thing is more a UX flow question: allow users to enter email password on the page (easier to understand maybe, but e.g. bad for password managers) (this is basically what bsky does) or redirect them to a pds and do the signup there with prompt=create 171 171 + byarielm — 2:13 PM 172 172 + Thanks! So to summarize and check my knowledge: Skyware is kinda the base code you need for a labeler, Ozone provides UI and some extra db/moderation stuff, and Osprey is more pattern/rule formatting for labellers? Could you use Osprey without Ozone? 173 173 + verdverm.com — 2:13 PM 174 174 + If you want to talk perm'd data, I'd be happy to join, I'm deeply involved 175 175 + 176 176 + also, best pronunciation of verdverm yet! 177 177 + flo.bit 178 178 + is now a speaker. — 2:13 PM 179 179 + n8 — 2:13 PM 180 180 + yes i'd love to learn what you're up to! 181 181 + verdverm.com — 2:14 PM 182 182 + I can more likely answer other's questions, the ones I have myself are open and as yet unanswered 183 183 + sorryforpartyrock.ing 184 184 + 185 185 + — 2:18 PM 186 186 + Hi all ! 187 187 + n8 — 2:18 PM 188 188 + 👋 189 189 + byarielm — 2:18 PM 190 190 + And it's only showing recently active sessions lol 🥲 so not really useful 191 191 + evan.jarrett.net 192 192 + 193 193 + — 2:19 PM 194 194 + yeah for client confidential long standing oauth, they don't show up all the time 195 195 + sorryforpartyrock.ing 196 196 + 197 197 + — 2:20 PM 198 198 + oh that's weird 199 199 + i thought theyd only show valid sessions and those would be valid 200 200 + byarielm — 2:20 PM 201 201 + the only one showing in mine rn is smoke signal (just went to the site a few min ago) + a list of device logins 202 202 + psingletary.com 203 203 + 204 204 + — 2:20 PM 205 205 + hi, i am a patrick 206 206 + pdewey.com — 2:21 PM 207 207 + hi i am also a patrick 208 208 + evan.jarrett.net 209 209 + 210 210 + — 2:21 PM 211 211 + i'm thinking about it instead of app-passwords 212 212 + thisismissem.social — 2:21 PM 213 213 + osprey is a automation and data analysis tool — it receives the events from the a source, e.g., the firehose, and then based on the rules your stored (a DSL) it can then apply those to data as it comes in, it's very similar in that way to SQRL. Osprey also stores all the events in a big database, which then allows you to write a new rules and investigate patterns 214 214 + verdverm.com — 2:21 PM 215 215 + apikey like thing? 216 216 + geesawra — 2:21 PM 217 217 + I am not Patrick sadly 218 218 + psingletary.com 219 219 + 220 220 + — 2:22 PM 221 221 + with patrick powers combined we are patrick² 222 222 + verdverm.com — 2:22 PM 223 223 + My Permissioned PDS POC has most of an APIKey implementation, which is not that much overall 224 224 + psingletary.com 225 225 + 226 226 + — 2:22 PM 227 227 + you can be an honorary one for this hangout 228 228 + evan.jarrett.net 229 229 + 230 230 + — 2:22 PM 231 231 + my thought is basically implementing a custom apikey that is just linked to your web logged in oauth session. rather than using app-passwords. its a pain because its basically just app-password but now tied to our oauth. 232 232 + @quillmatiq.com — 2:23 PM 233 233 + 👋🏼 234 234 + verdverm.com — 2:25 PM 235 235 + apikeys need a table in the db and all the other CRUD/LEX to go with them, but mechanically in the PDS stores (actor notably) logic, the auth method is largely abstracted at that point 236 236 + same.supply — 2:25 PM 237 237 + https://github.com/mary-ext/atcute/tree/trunk/packages/oauth/node-client-public-example 238 238 + thisismissem.social — 2:26 PM 239 239 + there's examples in bsky cookbook 240 240 + https://github.com/bluesky-social/cookbook/ 241 241 + evan.jarrett.net 242 242 + 243 243 + — 2:26 PM 244 244 + @n8 how about your adventures in relays 245 245 + thisismissem.social — 2:26 PM 246 246 + the IETF account management draft: https://github.com/ThisIsMissEm/draft-oauth-authorization-management-uri 247 247 + sharpie — 2:26 PM 248 248 + hi chat 249 249 + thisismissem.social — 2:27 PM 250 250 + yeah, your application could certainly do that for your API 251 251 + pdewey.com — 2:27 PM 252 252 + "the nostr guy" is a great way to describe that 253 253 + psingletary.com 254 254 + 255 255 + — 2:28 PM 256 256 + i thought they were also a svelte person 257 257 + flo.bit — 2:28 PM 258 258 + lets not give svelte people a bad rep 😛 259 259 + same.supply — 2:28 PM 260 260 + as a svelte person, the nostr guys can claim him 261 261 + @quillmatiq.com — 2:29 PM 262 262 + oops 263 263 + psingletary.com 264 264 + 265 265 + — 2:29 PM 266 266 + he they proclaiied it heavily in their bio... cowshrug 267 267 + Image 268 268 + verdverm.com — 2:29 PM 269 269 + I was in the trenches on HN for "the nostr guy" post, good place to see non-atproto perspectives 270 270 + @quillmatiq.com — 2:30 PM 271 271 + always nervous to hit HN -- worth going thru those comments? 272 272 + thisismissem.social — 2:30 PM 273 273 + https://atproto.com/guides/the-at-stack 274 274 + dane 275 275 + 276 276 + — 2:30 PM 277 277 + here's collectiondir for anyone wondering https://github.com/bluesky-social/indigo/tree/main/cmd/collectiondir 278 278 + evan.jarrett.net 279 279 + 280 280 + — 2:30 PM 281 281 + https://atcr.io/r/zzstoatzz.io/collectiondir 282 282 + ATCR 283 283 + zzstoatzz.io/collectiondir - ATCR 284 284 + collectiondir 285 285 + zzstoatzz.io/collectiondir - ATCR 286 286 + verdverm.com — 2:31 PM 287 287 + yeah, I hang out in the comments, my spicer atproto takes happen over there too (not so much this post) 288 288 + sorryforpartyrock.ing 289 289 + 290 290 + — 2:31 PM 291 291 + that's so much data 292 292 + n8 — 2:31 PM 293 293 + @Bryan Guffey idk if that didn't go thru, i tried to approve your request to speak! 294 294 + 295 295 + pls try again maybe? 296 296 + @quillmatiq.com — 2:32 PM 297 297 + Image 298 298 + sharpie — 2:32 PM 299 299 + i left twitter, you couldn't pay me to read nostr comments 300 300 + that's where all the crypto apologists went :( 301 301 + geesawra — 2:32 PM 302 302 + no we're also on bsky 303 303 + but we have less extreme views and are cool all around 304 304 + @quillmatiq.com — 2:32 PM 305 305 + not yet 🙂 306 306 + sorryforpartyrock.ing 307 307 + 308 308 + — 2:32 PM 309 309 + i used to look at using farcaster 310 310 + same.supply — 2:32 PM 311 311 + dozens! 312 312 + @quillmatiq.com — 2:32 PM 313 313 + but i see where you're going with it 314 314 + verdverm.com — 2:33 PM 315 315 + I'm probably one of the reasons this meme exists 🙈 🙊 🙉 316 316 + Bryan Guffey 317 317 + is now a speaker. — 2:33 PM 318 318 + sharpie — 2:33 PM 319 319 + crypto infrastructure isn't bad; the people using it unethically are tho 320 320 + geesawra — 2:33 PM 321 321 + agreed 322 322 + infra also sucks tho 323 323 + most of the times 324 324 + sharpie — 2:33 PM 325 325 + i like the concept 326 326 + but the implimentation is like everything else in tech 327 327 + good in an ideal world and not the real one 328 328 + sorryforpartyrock.ing 329 329 + 330 330 + — 2:34 PM 331 331 + big infrastructure is fun 332 332 + has real effects tho 🥲 333 333 + verdverm.com — 2:34 PM 334 334 + my atmunge project produces a table with every account and collections, it's very easy to build that index, one request per account 335 335 + geesawra — 2:34 PM 336 336 + you either end up being slow or centralized, no middle ground 337 337 + n8 — 2:34 PM 338 338 + hell yea! 339 339 + sorryforpartyrock.ing 340 340 + 341 341 + — 2:34 PM 342 342 + centralised 👀 343 343 + byarielm — 2:35 PM 344 344 + i'll be there! 🙂 (at the con lol) 345 345 + dane 346 346 + 347 347 + — 2:35 PM 348 348 + hooray for meetups 349 349 + sorryforpartyrock.ing 350 350 + 351 351 + — 2:35 PM 352 352 + i wish i was in the eu rn 353 353 + i'd go 354 354 + geesawra — 2:35 PM 355 355 + as a non-native speaker it's so confusing haha 356 356 + byarielm — 2:35 PM 357 357 + nobody come to buffalo u might get trapped 358 358 + psingletary.com 359 359 + 360 360 + — 2:35 PM 361 361 + Attending ATmosphereConf? 362 362 + IRL 363 363 + 364 364 + 4 votes 365 365 + 33% 366 366 + Virtual 367 367 + 368 368 + 5 votes 369 369 + 42% 370 370 + What's that? 371 371 + 372 372 + 2 votes 373 373 + 17% 374 374 + Nah 375 375 + 376 376 + 1 vote 377 377 + 8% 378 378 + 379 379 + 12 votes 380 380 + Poll closed 381 381 + sharpie — 2:36 PM 382 382 + i'm hosting an atproto table at my local event next week 383 383 + sorryforpartyrock.ing 384 384 + 385 385 + — 2:36 PM 386 386 + what's atmosphereconf? 387 387 + sharpie — 2:36 PM 388 388 + i don't think that's news worthy tho lmao 389 389 + sorryforpartyrock.ing 390 390 + 391 391 + — 2:36 PM 392 392 + for those of us who don't know 393 393 + geesawra — 2:36 PM 394 394 + needs a "nah but i wish" 395 395 + evan.jarrett.net 396 396 + 397 397 + — 2:36 PM 398 398 + https://atmosphereconf.org/ 399 399 + ATmosphereConf 2026 400 400 + ATmosphereConf is the global atproto community conference. Join us in Vancouver, Canada, March 26th - 29th, 2026. 401 401 + Image 402 402 + psingletary.com 403 403 + 404 404 + — 2:36 PM 405 405 + 2nd 406 406 + thisismissem.social — 2:36 PM 407 407 + There's an interesting thread in here about infra, apps, and moderation: https://bsky.app/profile/thisismissem.social/post/3mfg3fc2jwc24 408 408 + sharpie — 2:36 PM 409 409 + i'll be attending atmosphere virtually; don't feel comfortable leaving the US rn 410 410 + have fun tho !! 411 411 + geesawra — 2:37 PM 412 412 + atmosphereconf EU when 413 413 + psingletary.com 414 414 + 415 415 + — 2:37 PM 416 416 + same, if i leave i have to prepare that I can't come back 417 417 + dane 418 418 + 419 419 + — 2:37 PM 420 420 + next one should be eu 421 421 + i hope 422 422 + sorryforpartyrock.ing 423 423 + 424 424 + — 2:37 PM 425 425 + i'll be going up with eli i guess 426 426 + psingletary.com 427 427 + 428 428 + — 2:37 PM 429 429 + aren't you riding the train? 430 430 + sharpie — 2:37 PM 431 431 + yeah, unfortunately i can't take that risk rn. important things i'm working on locally 432 432 + Bryan Guffey 433 433 + 434 434 + — 2:38 PM 435 435 + run your own hahah 436 436 + dane 437 437 + 438 438 + — 2:38 PM 439 439 + logical next step i guess, we've had us then canada 440 440 + sorryforpartyrock.ing 441 441 + 442 442 + — 2:38 PM 443 443 + the train leaves at 6pm 444 444 + geesawra — 2:38 PM 445 445 + i want to put dotfiles on pdses 446 446 + ultimate nerd snipe 447 447 + @quillmatiq.com — 2:38 PM 448 448 + ✨ e v e r y t h i n g a c c o u n t ✨ 449 449 + geesawra — 2:38 PM 450 450 + counterpoint: do put them in your pds, make security engineers happy 451 451 + verdverm.com — 2:38 PM 452 452 + Why atproto over git for dotfiles? 453 453 + psingletary.com 454 454 + 455 455 + — 2:38 PM 456 456 + <-bitwarden stan 457 457 + geesawra — 2:38 PM 458 458 + fun! 459 459 + sharpie — 2:39 PM 460 460 + ohhhh i saw that 461 461 + evan.jarrett.net 462 462 + 463 463 + — 2:39 PM 464 464 + they mean if the dotfiles themselves could be encrypted/private 465 465 + thisismissem.social — 2:39 PM 466 466 + https://bsky.app/profile/why.bsky.team/post/3mfhon6ss4k2g 467 467 + same.supply — 2:39 PM 468 468 + I mean they're probably public on git (dotfiles) 469 469 + sorryforpartyrock.ing 470 470 + 471 471 + — 2:39 PM 472 472 + what about people who dont have iphones? 473 473 + geesawra — 2:39 PM 474 474 + android could do something like that 475 475 + sharpie 476 476 + is now a speaker. — 2:39 PM 477 477 + sorryforpartyrock.ing 478 478 + 479 479 + — 2:39 PM 480 480 + i have an galaxy s4 i use often 481 481 + psingletary.com 482 482 + 483 483 + — 2:40 PM 484 484 + S FOUR! 485 485 + geesawra — 2:40 PM 486 486 + should rotation keys be offered in bsky's settings? like under an "advance" use 487 487 + sorryforpartyrock.ing 488 488 + 489 489 + — 2:40 PM 490 490 + it's my main phone 🙉 491 491 + sorryforpartyrock.ing 492 492 + 493 493 + — 2:40 PM 494 494 + no 495 495 + well 496 496 + geesawra — 2:40 PM 497 497 + huh why? 498 498 + same.supply — 2:40 PM 499 499 + imo they should never be on an apps website 500 500 + dane 501 501 + 502 502 + — 2:40 PM 503 503 + i think so at least, i don't think it should be a separate thing 504 504 + sorryforpartyrock.ing 505 505 + 506 506 + — 2:41 PM 507 507 + i don't think it would end well, people would lose them quite often 508 508 + same.supply — 2:41 PM 509 509 + it should be on your pds host website 510 510 + sorryforpartyrock.ing 511 511 + 512 512 + — 2:41 PM 513 513 + i've made multiple and then immediately lost them 514 514 + @quillmatiq.com — 2:41 PM 515 515 + great explanation 516 516 + psingletary.com 517 517 + 518 518 + — 2:41 PM 519 519 + forwarding address 520 520 + sorryforpartyrock.ing 521 521 + 522 522 + — 2:41 PM 523 523 + bailey has mine so we good now 524 524 + @taurean.bryant.land — 2:41 PM 525 525 + really good, super useful thanks 526 526 + geesawra — 2:41 PM 527 527 + that's why i think it should be under an advanced menu, it's like, if you want to use them you can 528 528 + otherwise i feel they'll always be reserved for nerds like me 529 529 + sorryforpartyrock.ing 530 530 + 531 531 + — 2:42 PM 532 532 + i think they'd be too visible and associated with bsky 533 533 + i get what you mean tho 534 534 + if they get the secure enclave thing in-app i think that would be nice tho 535 535 + geesawra — 2:42 PM 536 536 + yeah exactly 537 537 + @taurean.bryant.land — 2:42 PM 538 538 + yeah, thats great too. like showing ID to fill out a change of address form with the post office 539 539 + sorryforpartyrock.ing 540 540 + 541 541 + — 2:42 PM 542 542 + theyre like 2fa backup keys 543 543 + geesawra — 2:42 PM 544 544 + who's building ledgers for plc 545 545 + sorryforpartyrock.ing 546 546 + 547 547 + — 2:42 PM 548 548 + ID is a little better i think 549 549 + you can't lose a fingerprint under normal circumstances but you can an ID 550 550 + psingletary.com 551 551 + 552 552 + — 2:43 PM 553 553 + lol let's not worry about semantics in favor of explaining a concept 554 554 + sharpie — 2:43 PM 555 555 + i tried explaining the public v private components and users usually get confused 556 556 + so i don't explain it 557 557 + geesawra — 2:44 PM 558 558 + asymmetric crypto is hard 559 559 + evan.jarrett.net 560 560 + 561 561 + — 2:44 PM 562 562 + i like the skeleton key concept even though thats also not quite accurate 563 563 + sorryforpartyrock.ing 564 564 + 565 565 + — 2:44 PM 566 566 + i think streamplace does it well enough 567 567 + @taurean.bryant.land — 2:44 PM 568 568 + 100%. maybe the only thing that could be useful is the directory stores the ability to verify your passport or key or whatever. 569 569 + psingletary.com 570 570 + 571 571 + — 2:44 PM 572 572 + bff heart necklaces 573 573 + sorryforpartyrock.ing 574 574 + 575 575 + — 2:44 PM 576 576 + a public key is associated with a stream key which is like a password 577 577 + thisismissem.social — 2:44 PM 578 578 + yeah, it is a tricky one, I'd just say you have a credential that you can use to manage your identity in the cloud, and when you create that credential, you register it with your identity in the cloud 579 579 + byarielm — 2:44 PM 580 580 + it's like lock-and-key matching but...more like fingerprints that are actually unique lol. i like your explanation sharpie, good for non tech users. 581 581 + psingletary.com 582 582 + 583 583 + — 2:44 PM 584 584 + one person has half of the necklace, you have the other 585 585 + sharpie — 2:45 PM 586 586 + yeah honestly, getting people to tune in long enough to even understand what the PLC is; is very hard 587 587 + i can hear people's eyes glazing over lmao 588 588 + geesawra — 2:45 PM 589 589 + what if the necklace was split in n parts and you could reconstruct the full necklace with m parts 👀 590 590 + @taurean.bryant.land — 2:45 PM 591 591 + I appreciate you @sharpie! we need more people grounding this to focus on people instead of tech 592 592 + sorryforpartyrock.ing 593 593 + 594 594 + — 2:46 PM 595 595 + only thing i ever use AI for is to explain big systems i'm not familiar with already 596 596 + sharpie 597 597 + is now a speaker. — 2:46 PM 598 598 + psingletary.com 599 599 + 600 600 + — 2:47 PM 601 601 + fig? 602 602 + sorryforpartyrock.ing 603 603 + 604 604 + — 2:47 PM 605 605 + code should be maintainable and that includes me either writing or reviewing code, ideally both 606 606 + @quillmatiq.com — 2:47 PM 607 607 + "if elon shows up..." 608 608 + @taurean.bryant.land — 2:47 PM 609 609 + I was just going to say "Eloned" lol 610 610 + @quillmatiq.com — 2:47 PM 611 611 + I ususally say "The Worst Person You Know" but they get the hint 612 612 + psingletary.com 613 613 + 614 614 + — 2:47 PM 615 615 + @phil fyi /nick fig (aka:[phil]) 616 616 + phil — 2:47 PM 617 617 + ❤️ https://updates.microcosm.blue/3lz7nwvh4zc2u 618 618 + Adversarial PLC directory migration - microcosm 619 619 + Contingency planning: what's our credible failover for the centralized underpinning of ATProto identities? 620 620 + Adversarial PLC directory migration - microcosm 621 621 + @taurean.bryant.land — 2:48 PM 622 622 + ‼️‼️‼️ 623 623 + animations would actually be amazing, the video that Sharpie did has already been so valuable, this would be great. 624 624 + @negritosuave.blacksky.app — 2:48 PM 625 625 + There’s a video? 626 626 + sorryforpartyrock.ing 627 627 + 628 628 + — 2:49 PM 629 629 + oh i'm awful with UI 630 630 + ui/ux i am absolute trash at 631 631 + geesawra — 2:49 PM 632 632 + UI is hard 633 633 + evan.jarrett.net 634 634 + 635 635 + — 2:49 PM 636 636 + I think we need a client facing name for the PDS. just like "atproto account". its not very consumer friendly 637 637 + @taurean.bryant.land — 2:49 PM 638 638 + let me try and find it 639 639 + psingletary.com 640 640 + 641 641 + — 2:50 PM 642 642 + yes let's abandon meta 643 643 + they can burn for all i care 644 644 + @quillmatiq.com — 2:50 PM 645 645 + docpop did a fantastic job 646 646 + they didn't promote it well enough on the fedi 647 647 + / outside the fedi 648 648 + @quillmatiq.com — 2:51 PM 649 649 + ^ it's not a hope, it's a goal 650 650 + @taurean.bryant.land — 2:51 PM 651 651 + https://bsky.app/profile/sharpiepls.com/post/3lw3hcuojck2u 652 652 + 653 653 + dapurplesharpie (@sharpiepls.com) 654 654 + Entry 31 - Migrating from the Bluesky PDS to the #BlackSky PDS with PDS MOOver #SharpieVLOG 655 655 + Reposts 656 656 + 147 657 657 + Likes 658 658 + 463 659 659 + Quotes 660 660 + 273 661 661 + 662 662 + Bluesky•8/10/25, 6:39 PM 663 663 + thisismissem.social — 2:51 PM 664 664 + https://bsky.app/profile/heika.dog/post/3mbfpu24vac2t 665 665 + flo.bit — 2:51 PM 666 666 + we also need a internet-handle website that shows the popular apps you can sign in with your atproto account, etc 667 667 + sharpie — 2:52 PM 668 668 + a community needs EVERYBODY 669 669 + n8 — 2:52 PM 670 670 + (i have a WIP for this https://waow.tech/ , not even close to all) 671 671 + flo.bit — 2:52 PM 672 672 + kinda like https://internethandle.org/ + https://atproto.brussels/atproto-apps but with better ux 673 673 + psingletary.com 674 674 + 675 675 + — 2:52 PM 676 676 + we might need to figure out how to move away from discord to make this more open, maybe streamplace with vdo.ninja, but audio routing always ends up being the nightmare 677 677 + geesawra — 2:53 PM 678 678 + there's a severe lack of specs in atmosphere 679 679 + sharpie — 2:53 PM 680 680 + ventrillo or mumble? 681 681 + thisismissem.social — 2:53 PM 682 682 + what do you mean by this? at a lexicon level? 683 683 + geesawra — 2:53 PM 684 684 + even at lower levels, i feel like there's an adequate amount of documentation but not enough rigorous spec 685 685 + psingletary.com 686 686 + 687 687 + — 2:54 PM 688 688 + yeah we are up to 12 viewers on @sorryforpartyrock.ing stream 689 689 + sharpie — 2:54 PM 690 690 + where are the recordings for the calls stored? 691 691 + @quillmatiq.com — 2:54 PM 692 692 + we can also start sans lexicon and move when we're ready, doesn't need to be a blocker 693 693 + psingletary.com 694 694 + 695 695 + — 2:54 PM 696 696 + we're not unless @n8 is recording 697 697 + geesawra — 2:54 PM 698 698 + mind you i might not've looked hard enough! but stuff like a "pds spec" mostly lives in the reference implementation not in writing 699 699 + sorryforpartyrock.ing 700 700 + 701 701 + — 2:54 PM 702 702 + why would we store this haha 703 703 + psingletary.com 704 704 + 705 705 + — 2:54 PM 706 706 + having the conversation be ephemeral is probably better though 707 707 + sorryforpartyrock.ing 708 708 + 709 709 + — 2:55 PM 710 710 + yeah that's my thought 711 711 + n8 — 2:55 PM 712 712 + yea im taking notes but haven't explored recording yet (maybe streamplace vods down the road but only if ppl actually want that) 713 713 + psingletary.com 714 714 + 715 715 + — 2:55 PM 716 716 + we kind of wander all over and that fine for participation but terrible for watchability 717 717 + geesawra — 2:55 PM 718 718 + yes I agree 719 719 + evan.jarrett.net 720 720 + 721 721 + — 2:55 PM 722 722 + i was expecting to see the com.atproto lexicons/xrpc on the new atproto site.... 723 723 + was kinda disappointed when they weren't there 724 724 + sorryforpartyrock.ing 725 725 + 726 726 + — 2:56 PM 727 727 + oh we do that 728 728 + yeah 729 729 + geesawra — 2:56 PM 730 730 + i guess my concept of spec is more tied to the conventional "this is a list of endpoints" rather than "here's xrpc and lexicons, go nuts" 731 731 + sorryforpartyrock.ing 732 732 + 733 733 + — 2:56 PM 734 734 + it's so funny 735 735 + thoth — 2:56 PM 736 736 + I'm still super concerned about the blocking repost suppression bug. 737 737 + geesawra — 2:56 PM 738 738 + yesss the deprecated tag is vital 739 739 + verdverm.com — 2:57 PM 740 740 + I already have a tool to do this, someone just needs to write the templates for markdown or whatever format you want 741 741 + @taurean.bryant.land — 2:57 PM 742 742 + Lexicon Garden is great: https://lexicon.garden/ 743 743 + Lexicon Garden 744 744 + Lexicon Garden - ATProtocol Lexicon Browser 745 745 + Browse and discover ATProtocol lexicon schema definitions 746 746 + Image 747 747 + for lexicon documentation 748 748 + sorryforpartyrock.ing 749 749 + 750 750 + — 2:57 PM 751 751 + how do you generate good documentation though? 752 752 + that's impractical i think 753 753 + i'm a big fan of self documenting code 754 754 + psingletary.com 755 755 + 756 756 + — 2:58 PM 757 757 + lol @sorryforpartyrock.ing you should let people know who you are 758 758 + thoth — 2:58 PM 759 759 + I mean the process described is "generate a skeleton from the lexicon and fill it in" 760 760 + sorryforpartyrock.ing 761 761 + 762 762 + — 2:58 PM 763 763 + no why? 764 764 + thoth — 2:58 PM 765 765 + But it'd probably be better to have a synthesis thing that generates docs and compiled lexicons from an easier-to-write source code description. 766 766 + So the docs and code can't get out of sync so easily. 767 767 + (and you don't have to do a weird dance when the record context changes) 768 768 + sorryforpartyrock.ing 769 769 + 770 770 + — 2:59 PM 771 771 + https://bsky-docs.pages.dev/ isn't too bad 772 772 + My Docs 773 773 + Bluesky & AT Protocol Documentation 774 774 + Unofficial documentation for Bluesky and the AT Protocol, including auto-generated lexicon references. 775 775 + oh those are completely different 776 776 + i think separate doc pages that tell you how to use things/full flow 777 777 + that's fine 778 778 + verdverm.com — 3:00 PM 779 779 + I have a tool that generally, and at scale, allows one to data + templates -> lots of files / dirs. Also allows you to write the extra, and also regen the gen'd parts if they change, without breaking things 780 780 + sorryforpartyrock.ing 781 781 + 782 782 + — 3:00 PM 783 783 + i love docs.bsky.app 784 784 + its so bad 785 785 + n8 — 3:00 PM 786 786 + examples/concepts/api-ref, i think we need all 3, some will be app level 787 787 + psingletary.com 788 788 + 789 789 + — 3:01 PM 790 790 + halfway mark 791 791 + sorryforpartyrock.ing 792 792 + 793 793 + — 3:01 PM 794 794 + yeah this is kind of what i do already 795 795 + byarielm — 3:01 PM 796 796 + we also need the "for babies"/"for newbs" docs 🥲 with all the caveats so i can learn more 797 797 + n8 — 3:01 PM 798 798 + "Introduction"/"Getting Started" 799 799 + sorryforpartyrock.ing 800 800 + 801 801 + — 3:01 PM 802 802 + we need more of a concepts thing 803 803 + geesawra — 3:02 PM 804 804 + the go documentation does this well, you have examples/docs/howtos plus reference in a single page 805 805 + Jack Valinsky — 3:02 PM 806 806 + *taking notes for my own projects wrt. docs 😬 807 807 + thisismissem.social — 3:02 PM 808 808 + yeah, this is the same as lexicon.garden really: a html rendering of the lexicon, but it's not prose 809 809 + evan.jarrett.net 810 810 + 811 811 + — 3:02 PM 812 812 + i tried to write documentation and then it got outdated like 3 commits later. 813 813 + sorryforpartyrock.ing 814 814 + 815 815 + — 3:02 PM 816 816 + yeah that's fine, it's not trying to be anything more 817 817 + thoth — 3:02 PM 818 818 + RustDoc. Make it like RustDoc. This is the way. 819 819 + sorryforpartyrock.ing 820 820 + 821 821 + — 3:02 PM 822 822 + you get it 823 823 + Bryan Guffey 824 824 + 825 825 + — 3:02 PM 826 826 + OMG YES ARIEL WE SHOULD DO THIS TOGETHER 827 827 + sorryforpartyrock.ing 828 828 + 829 829 + — 3:02 PM 830 830 + rustdoc is so goated 831 831 + Bryan Guffey 832 832 + 833 833 + — 3:02 PM 834 834 + ELI5 835 835 + Bryan Guffey 836 836 + 837 837 + — 3:03 PM 838 838 + facts 839 839 + thoth — 3:03 PM 840 840 + I didn't think it was possible for docs generated from source annotations to be good (I mean, have you seen doxygen?) until I saw rustdoc 841 841 + sorryforpartyrock.ing 842 842 + 843 843 + — 3:03 PM 844 844 + i have a pretty shitty claude-written mini docs system i have locally 845 845 + geesawra — 3:03 PM 846 846 + rustdoc and godoc are the goats 847 847 + n8 — 3:03 PM 848 848 + yea they were pretty responsive on gh, got a lil typo thing merged in a couple hours 849 849 + byarielm — 3:03 PM 850 850 + thanks miss em! i do have some unanswered q's for what i've read so far, though it is MUCH better than what it was before. 851 851 + sorryforpartyrock.ing 852 852 + 853 853 + — 3:03 PM 854 854 + like it's pretty poorly written + llmisms but it works for understanding things 855 855 + thisismissem.social — 3:03 PM 856 856 + yeah, keep notes on what you're missing 857 857 + sorryforpartyrock.ing 858 858 + 859 859 + — 3:04 PM 860 860 + packed full of info is nice 861 861 + byarielm — 3:04 PM 862 862 + i started this lmao but haven't continued cuz (1) wanted to wait for the atproto example to move to tap, (2) busy. https://leaflet.pub/9ee7a6e8-d4bc-4a83-a794-165de5637222 863 863 + same.supply — 3:04 PM 864 864 + What I've seen work from good js libs is Getting Started/Quick Start + Important Concepts + API Type specification 865 865 + sorryforpartyrock.ing 866 866 + 867 867 + — 3:04 PM 868 868 + why would i? i'm not a docs writer 869 869 + claude isn't either, really 870 870 + Jack Valinsky — 3:05 PM 871 871 + an aside: mfw I realized that markdown files can have comments that are invisible to renderers but if feed to an LLM -> very easy prompt injection... maybe md format for LLMs was a mistake 872 872 + byarielm — 3:05 PM 873 873 + thinking i just need to sit down with the docs + margin annotations open. then do issues/PRs later. 874 874 + thisismissem.social — 3:05 PM 875 875 + https://github.com/bluesky-social/atproto-website/issues 876 876 + GitHub 877 877 + bluesky-social/atproto-website 878 878 + Contribute to bluesky-social/atproto-website development by creating an account on GitHub. 879 879 + Contribute to bluesky-social/atproto-website development by creating an account on GitHub. 880 880 + byarielm — 3:06 PM 881 881 + cuz it's hard to read and then pause to write an issue lol 882 882 + n8 — 3:06 PM 883 883 + can confirm, i find hard to step out my weeds and imagine how people interpret my docs 884 884 + sorryforpartyrock.ing 885 885 + 886 886 + — 3:07 PM 887 887 + i have to write user facing docs every so often and it's absolutely horrible 888 888 + Bryan Guffey 889 889 + 890 890 + — 3:07 PM 891 891 + why are you so mean to the user haha 892 892 + sorryforpartyrock.ing 893 893 + 894 894 + — 3:07 PM 895 895 + condolences to people who do 896 896 + Bryan Guffey 897 897 + 898 898 + — 3:07 PM 899 899 + we're the best 900 900 + thisismissem.social — 3:07 PM 901 901 + it's a certain skill! 902 902 + n8 — 3:08 PM 903 903 + https://leaflet.pub/66b63f56-f335-4d41-8e4a-0eca566cfb22 904 904 + notes 905 905 + asjdlfk 906 906 + notes 907 907 + thisismissem.social — 3:08 PM 908 908 + ooh! yes! annotate documentation with margin.at!! 909 909 + sorryforpartyrock.ing 910 910 + 911 911 + — 3:08 PM 912 912 + i'm hoping it's yjs 913 913 + ueberdosis has a quite nice yjs based sync server 914 914 + flo.bit 915 915 + is now a speaker. — 3:09 PM 916 916 + sorryforpartyrock.ing 917 917 + 918 918 + — 3:10 PM 919 919 + i don't want to build this 🥲 920 920 + it's usually too slow to do it on demand 921 921 + so you just have to do it in the background 922 922 + psingletary.com 923 923 + 924 924 + — 3:11 PM 925 925 + there's a reason why Graze hasn't tackled this yet 926 926 + geesawra — 3:11 PM 927 927 + sheesh 928 928 + sorryforpartyrock.ing 929 929 + 930 930 + — 3:11 PM 931 931 + for everyone 932 932 + evan.jarrett.net 933 933 + 934 934 + — 3:11 PM 935 935 + the following feed specifically the non-algorithmic one. becuase you have to show every post for every follower that everyone has 936 936 + sorryforpartyrock.ing 937 937 + 938 938 + — 3:11 PM 939 939 + with nyt i don't think it'll be too bad bc nyt is designed to be closed 940 940 + verdverm.com — 3:12 PM 941 941 + I'm working towards a set of those expensive feeds just for me, don't have to do it at scale :] 942 942 + psingletary.com 943 943 + 944 944 + — 3:12 PM 945 945 + @flo.bit you should reach out to graze team and community to work out a middle ground feed 946 946 + dane 947 947 + 948 948 + — 3:12 PM 949 949 + https://witchsky.app/profile/did:plc:vpkhqolt662uhesyj6nxm7ys/post/3m2wsfeab322b 🥲 950 950 + Witchsky 951 951 + Why (@why.bsky.team) 952 952 + Running the following feed is like half of our production workload, its kinda silly how much work it is to do that 953 953 + 954 954 + ↘️ quoting bryan newbold (@bnewbold.net): 955 955 + 956 956 + a great way to break up bsky appview implementation work would be implementing "Following" as a regular feedgen. 957 957 + 958 958 + one of the harder pieces in general high-req-rate full-network scale, ... 959 959 + same.supply — 3:12 PM 960 960 + is it following feed for posts or events? 961 961 + geesawra — 3:12 PM 962 962 + this is a certified "designing data-intensive applications" moment 963 963 + psingletary.com 964 964 + 965 965 + — 3:12 PM 966 966 + I, for one, am not a fan of following feeds, i feel they re-enforce parasocial relationships and dehumanizes accounts 967 967 + sorryforpartyrock.ing 968 968 + 969 969 + — 3:13 PM 970 970 + haha that's funny 971 971 + geesawra — 3:13 PM 972 972 + oh yeah! i logged in the other day and it said "feed building, come back later" 973 973 + sorryforpartyrock.ing 974 974 + 975 975 + — 3:13 PM 976 976 + im going to do that 977 977 + i think they only take like 5k people in account when doing following? 978 978 + i forget exactly 979 979 + yeah 980 980 + byarielm — 3:16 PM 981 981 + i think it's a post from jaz and it's mind hurting but also funny 982 982 + geesawra — 3:16 PM 983 983 + i think there's a post by someone on the team explaining that? 984 984 + sorryforpartyrock.ing 985 985 + 986 986 + — 3:16 PM 987 987 + i don't want to think about this 988 988 + geesawra — 3:16 PM 989 989 + maybe bryan? 990 990 + sorryforpartyrock.ing 991 991 + 992 992 + — 3:16 PM 993 993 + i'll have to do this eventually 994 994 + following feed, recommendations 995 995 + byarielm — 3:16 PM 996 996 + https://jazco.dev/2025/02/19/imperfection/ 997 997 + sorryforpartyrock.ing— 3:16 PM 998 998 + recommended topics 999 999 + 1000 1000 + geesawra — 3:17 PM 1001 1001 + ahhh yes, jaz 1002 1002 + same.supply — 3:17 PM 1003 1003 + I don't think many of the lexicons other than app.bsky.feed.posts would have those problems yet 1004 1004 + sorryforpartyrock.ing 1005 1005 + 1006 1006 + — 3:18 PM 1007 1007 + backfilling all of that will be a pain 1008 1008 + there's also the option of only taking in people you care about 1009 1009 + yeah 1010 1010 + evan.jarrett.net 1011 1011 + 1012 1012 + — 3:19 PM 1013 1013 + listrecordbycollection entered the chat 1014 1014 + psingletary.com 1015 1015 + 1016 1016 + — 3:19 PM 1017 1017 + this is me erasure! 1018 1018 + 1019 1019 + fortunately we have another patrick 1020 1020 + sorryforpartyrock.ing 1021 1021 + 1022 1022 + — 3:19 PM 1023 1023 + i just do it myself 1024 1024 + it's not hard 1025 1025 + you can just check all your lexicons on the relay to get most of them 1026 1026 + won't get all of them just by the nature of the relay 1027 1027 + geesawra — 3:20 PM 1028 1028 + til 1029 1029 + @taurean.bryant.land — 3:21 PM 1030 1030 + really good point 1031 1031 + Jack Valinsky — 3:21 PM 1032 1032 + huh that's good to know... 1033 1033 + @taurean.bryant.land — 3:22 PM 1034 1034 + a good middle ground is also using the followgraph as a follow recommendation system to build a new social graph. 1035 1035 + geesawra — 3:22 PM 1036 1036 + the good old "block, unblock to force unfollow" trick doesn't work around here! 1037 1037 + Jack Valinsky — 3:22 PM 1038 1038 + this is why we can't have nice things :/ 1039 1039 + geesawra — 3:23 PM 1040 1040 + that's incredibly antisocial 1041 1041 + n8 — 3:23 PM 1042 1042 + yes i am doing something like this at the moment, lots of interesting things to explore here 1043 1043 + verdverm.com — 3:24 PM 1044 1044 + why should block activity force some unfollow? or generalized, why does writing a record in one repo force a write in another repo? 1045 1045 + geesawra — 3:24 PM 1046 1046 + no way lmao 1047 1047 + Bryan Guffey 1048 1048 + 1049 1049 + — 3:25 PM 1050 1050 + yeah, this is the problem. i don't necessarily think ti should, but I do think people should be incontrol of blocking people they don;'t want to engage with. 1051 1051 + @quillmatiq.com — 3:25 PM 1052 1052 + we can 1053 1053 + Jack Valinsky — 3:25 PM 1054 1054 + yes 1055 1055 + same.supply — 3:25 PM 1056 1056 + yes 1057 1057 + psingletary.com 1058 1058 + 1059 1059 + — 3:25 PM 1060 1060 + yes we can hear 1061 1061 + @quillmatiq.com — 3:25 PM 1062 1062 + it says you're muted but you're not lol 1063 1063 + psingletary.com 1064 1064 + 1065 1065 + — 3:25 PM 1066 1066 + just disable video, 1067 1067 + sorryforpartyrock.ing 1068 1068 + 1069 1069 + — 3:25 PM 1070 1070 + uh oh 1071 1071 + Jack Valinsky — 3:25 PM 1072 1072 + I think your video is desynced from audio 1073 1073 + psingletary.com 1074 1074 + 1075 1075 + — 3:25 PM 1076 1076 + LOL 1077 1077 + evan.jarrett.net 1078 1078 + 1079 1079 + — 3:25 PM 1080 1080 + lol 1081 1081 + @quillmatiq.com — 3:25 PM 1082 1082 + oop 1083 1083 + pdewey.com — 3:26 PM 1084 1084 + rip 1085 1085 + psingletary.com 1086 1086 + 1087 1087 + — 3:26 PM 1088 1088 + THAT WASN'T ME 1089 1089 + @taurean.bryant.land — 3:26 PM 1090 1090 + cliffhanger 1091 1091 + geesawra — 3:26 PM 1092 1092 + f 1093 1093 + @quillmatiq.com — 3:26 PM 1094 1094 + cliffhanger 1095 1095 + Jack Valinsky — 3:26 PM 1096 1096 + pds.rip 1097 1097 + byarielm — 3:26 PM 1098 1098 + ded 1099 1099 + sorryforpartyrock.ing 1100 1100 + 1101 1101 + — 3:26 PM 1102 1102 + what did eli talk about 1103 1103 + @taurean.bryant.land — 3:26 PM 1104 1104 + get out of my head 1105 1105 + @quillmatiq.com — 3:26 PM 1106 1106 + TELL ME WHAT ELI SAID 1107 1107 + Jack Valinsky — 3:26 PM 1108 1108 + defederated 1109 1109 + sorryforpartyrock.ing 1110 1110 + 1111 1111 + — 3:26 PM 1112 1112 + it's probably s2pa/muxl 1113 1113 + ¯\_(ツ)_/¯ 1114 1114 + psingletary.com 1115 1115 + 1116 1116 + — 3:26 PM 1117 1117 + you think N8 just got discord age checked? 1118 1118 + geesawra — 3:26 PM 1119 1119 + oof smells like ddos 1120 1120 + psingletary.com 1121 1121 + 1122 1122 + — 3:26 PM 1123 1123 + OH MY 1124 1124 + n8 — 3:27 PM 1125 1125 + lmao 1126 1126 + my ISP blows 1127 1127 + on phone now 1128 1128 + geesawra — 3:27 PM 1129 1129 + f for your isp 1130 1130 + n8 1131 1131 + is now a speaker. — 3:27 PM 1132 1132 + psingletary.com 1133 1133 + 1134 1134 + — 3:28 PM 1135 1135 + plz no inbox 1136 1136 + ffs 1137 1137 + geesawra — 3:28 PM 1138 1138 + the fact that one could engineer a solution like this makes me hopeful for the future of atproto, the design feels infinitely extensible 1139 1139 + Jack Valinsky — 3:28 PM 1140 1140 + quick tl;dr why inbox bad 1141 1141 + geesawra — 3:29 PM 1142 1142 + patrick pick up the mic 1143 1143 + same.supply — 3:29 PM 1144 1144 + intents 👀 1145 1145 + psingletary.com 1146 1146 + 1147 1147 + — 3:29 PM 1148 1148 + notification overload it the biggest problem all electronic communicaitons have 1149 1149 + i only have capacity for bluesky 1150 1150 + the discourse forum is hard 1151 1151 + geesawra — 3:30 PM 1152 1152 + this is protocol level, it wont be necessarily showed to the user 1153 1153 + byarielm — 3:30 PM 1154 1154 + all notifications are muted forever everywhere 🙂 1155 1155 + psingletary.com 1156 1156 + 1157 1157 + — 3:30 PM 1158 1158 + I'm on discord under protest, once i get all th creators i support off of them i'm out 1159 1159 + I'm in 100 and have to rotate 1160 1160 + sorryforpartyrock.ing 1161 1161 + 1162 1162 + — 3:30 PM 1163 1163 + there's only one i'm active in 1164 1164 + geesawra — 3:30 PM 1165 1165 + @thisismissem.social you should bridge everything to matrix 👀 1166 1166 + byarielm — 3:30 PM 1167 1167 + i only have notifs on for tiny groups with low-frequency messages. 1168 1168 + alexwykoff — 3:31 PM 1169 1169 + doll made an irc that uses atproto auth 1170 1170 + Jack Valinsky — 3:31 PM 1171 1171 + zulip feels very academic for me 1172 1172 + psingletary.com 1173 1173 + 1174 1174 + — 3:31 PM 1175 1175 + i don;t think i can tell you how much i want roomy to succeed 1176 1176 + byarielm — 3:31 PM 1177 1177 + learning about discord folders ❤️ 1178 1178 + Image 1179 1179 + geesawra — 3:31 PM 1180 1180 + LMAO shots fired to matrix 1181 1181 + byarielm — 3:31 PM 1182 1182 + the atproto folder takes up the whole damn sidebar lmao 1183 1183 + geesawra — 3:31 PM 1184 1184 + working on it 1185 1185 + Jack Valinsky — 3:32 PM 1186 1186 + I'm reminded of that Beeper project + matric bridges 1187 1187 + same.supply — 3:32 PM 1188 1188 + roomy is a discourse replacement akchually 1189 1189 + (this is not an official statement) 1190 1190 + sorryforpartyrock.ing 1191 1191 + 1192 1192 + — 3:32 PM 1193 1193 + i feel like penny kinda puts me off a lil 1194 1194 + lmao 1195 1195 + alexwykoff — 3:32 PM 1196 1196 + it's here https://github.com/dollspace-gay/concord 1197 1197 + GitHub 1198 1198 + GitHub - dollspace-gay/concord 1199 1199 + Contribute to dollspace-gay/concord development by creating an account on GitHub. 1200 1200 + Contribute to dollspace-gay/concord development by creating an account on GitHub. 1201 1201 + geesawra — 3:33 PM 1202 1202 + straight up fire fr fr 1203 1203 + psingletary.com 1204 1204 + 1205 1205 + — 3:33 PM 1206 1206 + i put all public atproto discord links in the AT Proto : Social Events resources page 1207 1207 + psingletary.com 1208 1208 + 1209 1209 + — 3:33 PM 1210 1210 + where did you get that from? or is this a joke? 1211 1211 + michaelthatsit — 3:33 PM 1212 1212 + excellent usecase 1213 1213 + like the time limit on shorts in youtube 1214 1214 + psingletary.com 1215 1215 + 1216 1216 + — 3:35 PM 1217 1217 + this conversation has officially wandered all over 1218 1218 + same.supply — 3:35 PM 1219 1219 + half-joke... it's a long story 1220 1220 + psingletary.com 1221 1221 + 1222 1222 + — 3:35 PM 1223 1223 + thx for doing that @Bryan Guffey 1224 1224 + michaelthatsit — 3:35 PM 1225 1225 + I'm like super new to ATProto but I'm noticing bluesky handles OG a little differently? 1226 1226 + geesawra — 3:35 PM 1227 1227 + a podcast you can discourse with! 1228 1228 + pdewey.com — 3:35 PM 1229 1229 + lurkers rise up 1230 1230 + michaelthatsit — 3:36 PM 1231 1231 + Open graph 1232 1232 + psingletary.com 1233 1233 + 's poll Attending ATmosphereConf? has closed. — 3:36 PM 1234 1234 + Virtual 1235 1235 + Winning answer • 42% 1236 1236 + 1237 1237 + View Poll 1238 1238 + sorryforpartyrock.ing 1239 1239 + 1240 1240 + — 3:36 PM 1241 1241 + discord is weird 1242 1242 + they have their own OG server for bluesky afaik 1243 1243 + evan.jarrett.net 1244 1244 + 1245 1245 + — 3:36 PM 1246 1246 + they have been working on open graph fixes lately 1247 1247 + they serve 1k x 1k images 1248 1248 + flo.bit — 3:37 PM 1249 1249 + i'll also leave stage for now, so other people can talk 🙂 1250 1250 + michaelthatsit — 3:37 PM 1251 1251 + Oh no like I'm working on an semi-social fitness logger called fitsky.app, and I want them to be able to share their workouts with dynamically generated OG images 1252 1252 + geesawra — 3:37 PM 1253 1253 + soon (tm) 1254 1254 + michaelthatsit — 3:37 PM 1255 1255 + And it works on opengraph.dev but not bluesky 1256 1256 + flo.bit — 3:37 PM 1257 1257 + yeah bluesky og stuff is weird 1258 1258 + tynanpurdy.com — 3:37 PM 1259 1259 + Howdy 1260 1260 + evan.jarrett.net 1261 1261 + 1262 1262 + — 3:37 PM 1263 1263 + tangled has some, i have some for my project 1264 1264 + flo.bit — 3:37 PM 1265 1265 + depends on "solved" 😄 1266 1266 + probably cache ^^ 1267 1267 + geesawra — 3:38 PM 1268 1268 + private data will enable so many strava replacements, i can't wait 1269 1269 + michaelthatsit — 3:38 PM 1270 1270 + this guy 1271 1271 + https://ilo.so/bluesky-link-preview 1272 1272 + ilo.so 1273 1273 + Bluesky Link Previewer 1274 1274 + Preview how your link will look in a Bluesky post. 1275 1275 + Bluesky Link Previewer 1276 1276 + sri.xyz — 3:39 PM 1277 1277 + https://cardyb.bsky.app/v1/extract?url=https://atproto.com 1278 1278 + sorryforpartyrock.ing 1279 1279 + 1280 1280 + — 3:39 PM 1281 1281 + wait 1282 1282 + i have a rewrite of cardyb 1283 1283 + byarielm — 3:39 PM 1284 1284 + @sharpie how do you (if you do) explain the atmosphere/atprotocol to folks who maybe aren't even on bsky? 1285 1285 + michaelthatsit — 3:39 PM 1286 1286 + Lol I noticed that. I thought claude was hallucinating 1287 1287 + tynanpurdy.com — 3:39 PM 1288 1288 + What did you do to cardy A 1289 1289 + geesawra — 3:39 PM 1290 1290 + lol 1291 1291 + pdewey.com — 3:39 PM 1292 1292 + The first time I saw the cardyb user agent in my logs for arabica I was really confused 1293 1293 + sorryforpartyrock.ing 1294 1294 + 1295 1295 + — 3:39 PM 1296 1296 + https://github.com/espeon/gizo 1297 1297 + GitHub 1298 1298 + GitHub - espeon/gizo: A link card generator service 1299 1299 + A link card generator service. Contribute to espeon/gizo development by creating an account on GitHub. 1300 1300 + GitHub - espeon/gizo: A link card generator service 1301 1301 + Jack Valinsky — 3:39 PM 1302 1302 + I don't have anything but just thoughts to throw into the aether: did we talk about admin uis for PDS servers (bailey made a thread) ? 1303 1303 + 1304 1304 + or thinking about atproto based package managers (atpkgs, fair.pm, etc) -> what about a model weight registry for ex: roost.tools to download a model for moderation and audit it 1305 1305 + michaelthatsit — 3:39 PM 1306 1306 + still waiting on cardy c-z 1307 1307 + sharpie — 3:39 PM 1308 1308 + what platforms are they on and what is their familiarity with technology in general 1309 1309 + pdewey.com — 3:39 PM 1310 1310 + yeah I can talk about it, Its my service 1311 1311 + sorryforpartyrock.ing 1312 1312 + 1313 1313 + — 3:39 PM 1314 1314 + coffee? 1315 1315 + OH 1316 1316 + yeah that's epic 1317 1317 + pdewey.com 1318 1318 + is now a speaker. — 3:39 PM 1319 1319 + thisismissem.social — 3:40 PM 1320 1320 + there'll hopefully be news some time soon on things here too 1321 1321 + geesawra — 3:40 PM 1322 1322 + aeropress recipes on atproto??? 1323 1323 + byarielm — 3:40 PM 1324 1324 + perf, thanks. so you usually start with questions not explaining, that right? 1325 1325 + evan.jarrett.net 1326 1326 + 1327 1327 + — 3:40 PM 1328 1328 + https://tangled.org/did:plc:chqc2ockzmyvlrasfb66x64a/arabica 1329 1329 + Tangled 1330 1330 + arabica.social/arabica 1331 1331 + Coffee journaling on ATProto (alpha) 1332 1332 + arabica.social/arabica 1333 1333 + @quillmatiq.com — 3:40 PM 1334 1334 + ☕ 1335 1335 + evan.jarrett.net 1336 1336 + 1337 1337 + — 3:40 PM 1338 1338 + https://alpha.arabica.social/ 1339 1339 + Arabica 1340 1340 + Home - Arabica 1341 1341 + Track your coffee brewing journey. Built on AT Protocol, your data stays yours. 1342 1342 + michaelthatsit — 3:40 PM 1343 1343 + lol love it 1344 1344 + geesawra — 3:40 PM 1345 1345 + this is great 1346 1346 + Jack Valinsky — 3:40 PM 1347 1347 + https://developer.mozilla.org/en-US/docs/Web/HTTP/Reference/Status/418 🤭 1348 1348 + sharpie — 3:41 PM 1349 1349 + yeah. to understand their familiarity levels and then choosing an explaination based on that. you need to know who you're talking to and how they understand technology. 1350 1350 + n8 — 3:41 PM 1351 1351 + that’s sweet! 1352 1352 + psingletary.com 1353 1353 + 1354 1354 + — 3:41 PM 1355 1355 + https://alpha.arabica.social/ 1356 1356 + @taurean.bryant.land 1357 1357 + is now a speaker. — 3:41 PM 1358 1358 + tynanpurdy.com — 3:41 PM 1359 1359 + My go to is to throw a blento on the root 1360 1360 + sorryforpartyrock.ing 1361 1361 + 1362 1362 + — 3:41 PM 1363 1363 + yeah i have the same thing with nyt.one 1364 1364 + https://nyt.one/ 1365 1365 + flo.bit — 3:41 PM 1366 1366 + 😄 1367 1367 + sorryforpartyrock.ing 1368 1368 + 1369 1369 + — 3:41 PM 1370 1370 + inb4 "tanstack router app" embed 1371 1371 + sharpie — 3:41 PM 1372 1372 + shoutout to techs 1373 1373 + tynanpurdy.com — 3:41 PM 1374 1374 + Yoooo taurean 1375 1375 + byarielm — 3:41 PM 1376 1376 + thought so 🙂 would love to hear some different cases + what you'd say at some point! you are my comm inspo for atproto haha 1377 1377 + geesawra — 3:41 PM 1378 1378 + we need james hoffmann on this 1379 1379 + Jack Valinsky — 3:42 PM 1380 1380 + wafrn has some kind of ama/question mechanism so I mentioned questionable to them @thisismissem.social 1381 1381 + Bryan Guffey 1382 1382 + 1383 1383 + — 3:42 PM 1384 1384 + I'm gonna be that cat and set a weekly time for folks to come teach people atproto concepts 1385 1385 + thisismissem.social — 3:42 PM 1386 1386 + they know, I'm friendly with gabboman — they also know my plans beyond questionable.fyi 1387 1387 + sharpie — 3:42 PM 1388 1388 + this is like the time that someone posted about having username @ domain being their email too and i realized it was the tech 1389 1389 + psingletary.com 1390 1390 + 1391 1391 + — 3:42 PM 1392 1392 + not a coffee person, but fragrences is very interesting 1393 1393 + @taurean.bryant.land — 3:42 PM 1394 1394 + https://drydown.social/profile/taurean.bryant.land/house/3mcn34apccq2z 1395 1395 + Drydown - Fragrance Reviews 1396 1396 + Discover and track fragrances on Drydown. 1397 1397 + Drydown - Fragrance Reviews 1398 1398 + sorryforpartyrock.ing 1399 1399 + 1400 1400 + — 3:42 PM 1401 1401 + i love drydown haha 1402 1402 + has anyone checked out ouais' anki esque thing 1403 1403 + tynanpurdy.com — 3:42 PM 1404 1404 + Can the aromatic notes of a coffee review be the same lex as a fragrance review… 1405 1405 + thisismissem.social — 3:43 PM 1406 1406 + funny thing about mastodon handles: some folks think they can email the handle and nooope. 1407 1407 + sorryforpartyrock.ing 1408 1408 + 1409 1409 + — 3:43 PM 1410 1410 + *owais 1411 1411 + wow 1412 1412 + Jack Valinsky — 3:43 PM 1413 1413 + if not email why email shaped :/ 1414 1414 + thisismissem.social — 3:43 PM 1415 1415 + because webfinger, tagging @blaine.bsky.social 1416 1416 + sorryforpartyrock.ing 1417 1417 + 1418 1418 + — 3:44 PM 1419 1419 + webbed finger 1420 1420 + oh speaking of, i wonder how far that tailscale oauth sign in with bsky thing is going 1421 1421 + Jack Valinsky — 3:44 PM 1422 1422 + haha I vaguely knew that, and webfinger isn't even a part of the original AP spec (didn't define handle spec I think?) mastodon just did it b/c OStatus? 1423 1423 + psingletary.com 1424 1424 + 1425 1425 + — 3:44 PM 1426 1426 + GDIT these projects are already cards on semble 1427 1427 + thisismissem.social — 3:45 PM 1428 1428 + yeah, webfinger is a mastodon-ism which became defacto for the fediverse and the spec binding webfinger and activitypub was only written in the past 4 years 1429 1429 + sharpie — 3:45 PM 1430 1430 + this is actually super cool and i would love to talk about creator usecases @@taurean.bryant.land 1431 1431 + blaine.bsky.social — 3:45 PM 1432 1432 + funny thing about webfinger: google shipped support for it on gmail addresses in ~2009. The original intent was for it to work on email addresses / Thought experiment: maybe mastodon handles should be email addresses, too? 1433 1433 + @taurean.bryant.land — 3:45 PM 1434 1434 + would love to! 1435 1435 + thisismissem.social — 3:45 PM 1436 1436 + From 2024: https://swicg.github.io/activitypub-webfinger/ 1437 1437 + psingletary.com 1438 1438 + 1439 1439 + — 3:45 PM 1440 1440 + i'll get the scoop on @Hilary Baumann or @byarielm one of these days 1441 1441 + byarielm — 3:46 PM 1442 1442 + what's my scoop 1443 1443 + byarielm — 3:46 PM 1444 1444 + why scoop me 1445 1445 + geesawra — 3:46 PM 1446 1446 + are you ice cream 1447 1447 + same.supply — 3:46 PM 1448 1448 + YES 1449 1449 + byarielm — 3:46 PM 1450 1450 + no but i want some 1451 1451 + psingletary.com 1452 1452 + 1453 1453 + — 3:46 PM 1454 1454 + you and hilary finding sites before me 1455 1455 + blaine.bsky.social — 3:46 PM 1456 1456 + Webfinger didn't land in activitypub because standards-wonks-who-shall-not-be-named believe that you can't be "cool" unless you own your own domain name, and that personal identity should be only attached to web addresses. But that's a story for another time. 😉 1457 1457 + sharpie — 3:46 PM 1458 1458 + this becomes a convo about private data 1459 1459 + sorryforpartyrock.ing 1460 1460 + 1461 1461 + — 3:46 PM 1462 1462 + i think lexicon driven development is good but sometimes that's not possible 1463 1463 + flo.bit — 3:46 PM 1464 1464 + blento still doesnt have a lexicon 😅 1465 1465 + sorryforpartyrock.ing 1466 1466 + 1467 1467 + — 3:47 PM 1468 1468 + i think for e.g. providing secret keys 1469 1469 + geesawra — 3:47 PM 1470 1470 + every atproto convo eventually converges to private data 1471 1471 + sorryforpartyrock.ing 1472 1472 + 1473 1473 + — 3:47 PM 1474 1474 + it's not atproto but it's also fine 1475 1475 + evan.jarrett.net 1476 1476 + 1477 1477 + — 3:47 PM 1478 1478 + yeah i'm kinda in the middle of migrating my records and have to wait for everyone to slowly move over 1479 1479 + Jack Valinsky — 3:47 PM 1480 1480 + yeah there are weird semantics around the atproto design where techically you have an account on every possible service now and in the future in a sense depending on how people implement the client... instructions unclear follow Obama on tangled 1481 1481 + psingletary.com 1482 1482 + 1483 1483 + — 3:47 PM 1484 1484 + does it need one though? 1485 1485 + sri.xyz — 3:47 PM 1486 1486 + 🙁 1487 1487 + evan.jarrett.net 1488 1488 + 1489 1489 + — 3:47 PM 1490 1490 + called out oof 1491 1491 + sorryforpartyrock.ing 1492 1492 + 1493 1493 + — 3:48 PM 1494 1494 + thats so funny 1495 1495 + @taurean.bryant.land — 3:48 PM 1496 1496 + would love to show my latest fragrance and coffee reviews on blento btw 👀 1497 1497 + psingletary.com 1498 1498 + 1499 1499 + — 3:48 PM 1500 1500 + Not Everything Needs A Lexicon 1501 1501 + sharpie — 3:48 PM 1502 1502 + what is the fix for that? 1503 1503 + Jack Valinsky — 3:48 PM 1504 1504 + fig could just snipe people watching ufo feed 1505 1505 + flo.bit — 3:48 PM 1506 1506 + well ¯\_(ツ)_/¯ would probably be nice for other people to e.g. show blento cards somewhere else 1507 1507 + sorryforpartyrock.ing 1508 1508 + 1509 1509 + — 3:48 PM 1510 1510 + should i discontinue aktivi 1511 1511 + hrmm 1512 1512 + two minds about this et al 1513 1513 + flo.bit — 3:49 PM 1514 1514 + added to the list, but also feel free to submit a PR 🙂 https://github.com/flo-bit/blento 1515 1515 + Jack Valinsky — 3:49 PM 1516 1516 + @flo.bit tangled card for blento 🥺 👉 👈 1517 1517 + *maybe I should write a pr 🧐 1518 1518 + michaelthatsit — 3:50 PM 1519 1519 + is anyone going to the ATProto conference? 1520 1520 + geesawra — 3:50 PM 1521 1521 + thanks yall for arranging this 1522 1522 + @taurean.bryant.land — 3:50 PM 1523 1523 + Thanks for hosting this as always! 1524 1524 + same.supply 1525 1525 + is now a speaker. — 3:50 PM 1526 1526 + Bryan Guffey 1527 1527 + is now a speaker. — 3:50 PM 1528 1528 + Jack Valinsky — 3:50 PM 1529 1529 + oh right 1530 1530 + yeahhhh just a second 1531 1531 + thisismissem.social — 3:50 PM 1532 1532 + throwing you on the spot there 😂 1533 1533 + psingletary.com 1534 1534 + is now a speaker. — 3:50 PM 1535 1535 + same.supply — 3:50 PM 1536 1536 + https://bsky.app/profile/did:plc:2xau7wbgdq4phuou2ypwuen7/post/3l7esaih44s2s 1537 1537 + 1538 1538 + Tom Sherman (@tom.sherman.is) 1539 1539 + Seriously mild take: ATProto web clients should serve a <link rel="alternate"> containing an at:// URI 1540 1540 + 1541 1541 + I've just opened a PR to do this on Frontpage and atproto-browser. I'd for other clients to follow along with the convention! 1542 1542 + 1543 1543 + github.com/likeandscrib... 1544 1544 + 1545 1545 + Add AT URIs as alternate links by tom-sherman · Pull Request #172 · likeandscribe/frontpage 1546 1546 + I&#39;m trying to start a convention that allows tooling (eg. crawlers, browsers, extensions) to link the web representation of a record to it&#39;s canonical at:// URI. I see no better way to try ... 1547 1547 + Image 1548 1548 + 1549 1549 + Bluesky•10/25/24, 7:09 PM 1550 1550 + Jack Valinsky — 3:50 PM 1551 1551 + I've never used this interface for discord lol 1552 1552 + psingletary.com 1553 1553 + 1554 1554 + — 3:50 PM 1555 1555 + @same.supply it's really hard to hear you 1556 1556 + thisismissem.social — 3:51 PM 1557 1557 + <link rel=alternate> 1558 1558 + michaelthatsit — 3:51 PM 1559 1559 + plz do 1560 1560 + thisismissem.social — 3:51 PM 1561 1561 + that's what's being talked about, because then a web crawler can discover the atprotocol record for that page 1562 1562 + michaelthatsit — 3:51 PM 1563 1563 + ya boi needs office hours lol 1564 1564 + thisismissem.social — 3:51 PM 1565 1565 + there's maybe another microformat-y way to do that too 1566 1566 + thisismissem.social 1567 1567 + is now a speaker. — 3:52 PM 1568 1568 + sorryforpartyrock.ing 1569 1569 + 1570 1570 + — 3:52 PM 1571 1571 + i think you know enough to do this tbh 1572 1572 + more than me 1573 1573 + byarielm — 3:52 PM 1574 1574 + i'm da baby 1575 1575 + geesawra — 3:52 PM 1576 1576 + i love your songs and game consoles 1577 1577 + psingletary.com 1578 1578 + 1579 1579 + — 3:53 PM 1580 1580 + Let's speed round these statements so i can get on 1581 1581 + byarielm — 3:53 PM 1582 1582 + i thrive on chaos miss em lmao 1583 1583 + byarielm — 3:53 PM 1584 1584 + wat 1585 1585 + thisismissem.social — 3:53 PM 1586 1586 + Oh, specifically, the office hours I'd do would be public. i.e., I'm not giving 1:1 help 1587 1587 + If you want 1:1 help, hire me 🙂 1588 1588 + Bryan Guffey 1589 1589 + 1590 1590 + — 3:54 PM 1591 1591 + haha i wish i had the MONEYYYY 1592 1592 + thisismissem.social 1593 1593 + is now a speaker. — 3:54 PM 1594 1594 + thisismissem.social 1595 1595 + is now a speaker. — 3:55 PM 1596 1596 + michaelthatsit — 3:55 PM 1597 1597 + "f*ck you, pay me" - Socrates 1598 1598 + byarielm — 3:55 PM 1599 1599 + is ozone the only option for moderation workflow on labelers rn btw? looks like skyware doesn't and idk if mary plans it for the atcute setup 1600 1600 + n8 — 3:55 PM 1601 1601 + thank you for showing up ❤️ 1602 1602 + psingletary.com 1603 1603 + 1604 1604 + — 3:55 PM 1605 1605 + @thisismissem.social streamt o stream place and invite people with smokesignal 1606 1606 + evan.jarrett.net 1607 1607 + 1608 1608 + — 3:55 PM 1609 1609 + https://github.com/haileyok/osprey-for-atproto 1610 1610 + GitHub 1611 1611 + GitHub - haileyok/osprey-for-atproto: A high-performance safety rul... 1612 1612 + A high-performance safety rules engine for real-time event processing at scale - haileyok/osprey-for-atproto 1613 1613 + GitHub - haileyok/osprey-for-atproto: A high-performance safety rul... 1614 1614 + psingletary.com 1615 1615 + 1616 1616 + — 3:56 PM 1617 1617 + eli regularly gets tons of drive by 1618 1618 + Jack Valinsky — 3:56 PM 1619 1619 + I can try to talk about it, I think mine and @thisismissem.social 's thing is short to talk about if there is time 1620 1620 + geesawra — 3:57 PM 1621 1621 + eternal optimism by patrick 1622 1622 + thisismissem.social — 3:57 PM 1623 1623 + you started it, I just nerd sniped you, lol 1624 1624 + sharpie — 3:57 PM 1625 1625 + :o 1626 1626 + pdewey.com — 3:57 PM 1627 1627 + 👀 1628 1628 + sharpie — 3:57 PM 1629 1629 + WHAT 1630 1630 + HOW I GET IN THIS 1631 1631 + @quillmatiq.com — 3:57 PM 1632 1632 + love it 1633 1633 + thisismissem.social — 3:58 PM 1634 1634 + well, correction @Jack Valinsky you nerd sniped me so I nerd sniped you back 1635 1635 + byarielm — 3:58 PM 1636 1636 + so i shouldn't post the roomy link here? 1637 1637 + cuz i was gonna 1638 1638 + sharpie — 3:58 PM 1639 1639 + @psingletary.com 👀 1640 1640 + evan.jarrett.net 1641 1641 + 1642 1642 + — 3:58 PM 1643 1643 + not our super secret roomy chat 🤫 1644 1644 + sorryforpartyrock.ing 1645 1645 + 1646 1646 + — 3:58 PM 1647 1647 + ok bye 1648 1648 + Jack Valinsky 1649 1649 + is now a speaker. — 3:58 PM 1650 1650 + thisismissem.social 1651 1651 + is now a speaker. — 3:59 PM 1652 1652 + Jack Valinsky — 3:59 PM 1653 1653 + I can't unmute 1654 1654 + :/ 1655 1655 + thisismissem.social 1656 1656 + is now a speaker. — 3:59 PM 1657 1657 + michaelthatsit — 3:59 PM 1658 1658 + Honestly I'd love a youtube series on the ATProto just breaking down how it works. 1659 1659 + @taurean.bryant.land — 3:59 PM 1660 1660 + nailed it 1661 1661 + 🙌 1662 1662 + thisismissem.social — 3:59 PM 1663 1663 + production requires budget 🙁 1664 1664 + zeu — 3:59 PM 1665 1665 + I JUST GOT HEWW 1666 1666 + flo.bit — 3:59 PM 1667 1667 + 🙌 1668 1668 + pdewey.com — 3:59 PM 1669 1669 + thanks for hosting again! 1670 1670 + zeu — 3:59 PM 1671 1671 + NOOOO 1672 1672 + lmfao 1673 1673 + geesawra — 4:00 PM 1674 1674 + g2g to sleep 1675 1675 + bye folks! 1676 1676 + Bryan Guffey 1677 1677 + 1678 1678 + — 4:00 PM 1679 1679 + or do we wanna go to voice chat with no stage? 1680 1680 + Jack Valinsky 1681 1681 + is now a speaker. — 4:00 PM 1682 1682 + Bryan Guffey 1683 1683 + 1684 1684 + — 4:00 PM 1685 1685 + OHHH 1686 1686 + I have a LOT OF INFO FOR THTAT 1687 1687 + Jack Valinsky — 4:00 PM 1688 1688 + firefox :/ 1689 1689 + byarielm — 4:00 PM 1690 1690 + look we wanna hang and zeu is here now let us be 1691 1691 + psingletary.com 1692 1692 + 1693 1693 + — 4:00 PM 1694 1694 + Save it for next time the audience is alreayd kind of leaving 1695 1695 + Jack Valinsky — 4:00 PM 1696 1696 + yeah I just can't unmute 1697 1697 + pdewey.com — 4:00 PM 1698 1698 + also if anyone has feedback/suggestions for arabica feel free to shoot me a message on discord/bsky or file an issue 1699 1699 + byarielm — 4:01 PM 1700 1700 + huh 1701 1701 + Mia (parakeet.at) 1702 1702 + 1703 1703 + — 4:02 PM 1704 1704 + bi-directional nerdsniping 1705 1705 + Jack Valinsky — 4:02 PM 1706 1706 + I can't unmute 1707 1707 + firefox bad 1708 1708 + michaelthatsit — 4:02 PM 1709 1709 + love this platform for that lol 1710 1710 + Jack Valinsky — 4:03 PM 1711 1711 + pds-operator 1712 1712 + tynanpurdy.com — 4:03 PM 1713 1713 + Is this gonna be the impetus for record versioning lol 1714 1714 + Jack Valinsky — 4:03 PM 1715 1715 + ⁠pds-operators⁠ 1716 1716 + tynanpurdy.com — 4:04 PM 1717 1717 + And content licensing! 1718 1718 + Bryan Guffey 1719 1719 + 1720 1720 + — 4:04 PM 1721 1721 + coming to that too tynan 1722 1722 + Jack Valinsky — 4:04 PM 1723 1723 + there was debate about whether it should use site.standard or be strictly its own thing and I think b/c of stuff like versioning and tombstoning it should be its own thing 1724 1724 + n8 — 4:05 PM 1725 1725 + idk about other mod (patrick s) but i have to bounce at 4:15 central, just fyi 1726 1726 + thisismissem.social — 4:05 PM 1727 1727 + I think you'd have a standard.site for viewing but not for managing/publishing/accepting 1728 1728 + Jack Valinsky — 4:06 PM 1729 1729 + Forwarded 1730 1730 + another sensible option i think is to use the CID of the record as rkey whenever you create it and then make links back when you update. that way you get a content addressed history of the document and you just have another mutable record link to the currently active version 1731 1731 + #pds-operators • Friday 1732 1732 + nelind had this insight too 1733 1733 + the pds operator account signs it to so they have a record? 1734 1734 + nichoth — 4:07 PM 1735 1735 + sounds like legal challenge not code 1736 1736 + Jack Valinsky — 4:07 PM 1737 1737 + nick has done stuff on attestation 1738 1738 + thisismissem.social — 4:07 PM 1739 1739 + yeah, you can basically have the app sign the record that they write to your pds for the acceptance 1740 1740 + nichoth — 4:08 PM 1741 1741 + the attestation use case is a big opportunity imo 1742 1742 + Jack Valinsky — 4:09 PM 1743 1743 + idk if the word is "collection" but the the lexicon should have a <nsid>/region/policy 1744 1744 + 1745 1745 + sorry brain not lexiconing rn 1746 1746 + my apologies Bryan what is your at handle in case I have questions about potential legal gotchas with this idea? thanks 1747 1747 + Bryan Guffey 1748 1748 + 1749 1749 + — 4:10 PM 1750 1750 + @chaosgreml.in 1751 1751 + Jack Valinsky — 4:12 PM 1752 1752 + yeah we need to make it simple and straightforward to be informed b/c we want some random student to host a project and not be scared b/c of regulations that they are doing it wrong 1753 1753 + students are doing really cool stuff 1754 1754 + yeah! that'd be cool 1755 1755 + I will remember not to use firefox the next time for this hangout 😛 1756 1756 + flo.bit — 4:14 PM 1757 1757 + bye 🙂 1758 1758 + michaelthatsit — 4:15 PM 1759 1759 + enjoy! 1760 1760 + @taurean.bryant.land — 4:15 PM 1761 1761 + heading out, thanks yall! 1762 1762 + Jack Valinsky — 4:15 PM 1763 1763 + thanks for this 1764 1764 + byarielm — 4:15 PM 1765 1765 + yum 1766 1766 + oh no 1767 1767 + michaelthatsit — 4:15 PM 1768 1768 + enjoy more? 1769 1769 + byarielm — 4:15 PM 1770 1770 + byeee all! 1771 1771 + tynanpurdy.com — 4:15 PM 1772 1772 + Thanks yall! 1773 1773 + byarielm — 4:15 PM 1774 1774 + i'ma go eat too sry bryan 1775 1775 + tynanpurdy.com — 4:15 PM 1776 1776 + Have a good Sunday Bryan 1777 1777 + </discord chat>