experiments in a post-browser web
todo updates
+10
-4
+10
-4
TODO.md
+10
-4
TODO.md
···
200
200
We need a generalized system for feeds and time-series data.
201
201
It might be two different systems, based on analysis of the use-cases.
202
202
It might be a new system + the items system we already have, and using metadata.
203
+
The use-cases below cover a variety of feed-like needs.
203
204
204
205
Use-case: Querying web pages for time-series data
206
+
- Pattern: checking urls, collecting web data over time
205
207
- Eg a bank account balance, stock price, weather report
206
208
- Periodically load a web page in background hiddent window, and run a content script against it, returning a value and writing it to datastore (and metadata like page url, created/updated, custom metadata)
207
209
- Have a widget that shows latest value and its age
208
210
- Have a widget that shows graph of values over time
209
211
210
212
Use-case: Checking web page for ticket availability
213
+
- Pattern: checking urls for a one-time trigger
211
214
- Want to know when tickets go on sale for an event
212
215
- Periodically load a page in background, and run a content script against it
213
216
- Notify user when page change matched some criteria
214
217
215
218
Use-case: Pull data from external systems
219
+
- Pattern: harvesting 3rd party content feeds
216
220
- Eg RSS feeds or importing my posts/etc from other accounts - plays on Soundcloud, or thumbs-ups on Youtube, etc
217
221
- Imagine an extension which walks you through Oauth login for various systems, and then pulls their data into Peek via the system's API. Could be URLs with custom metadata (like Soundcloud plays) or non-URL data (like sports scores)
218
222
- Or someone could implement a feed reader extension focused on RSS reading
219
223
220
224
Use-case: System metrics
225
+
- Pattern: aggregating system metadata
221
226
- We need to be able to log and reflect all kinds of system information
222
227
- See the "Metadata, QS and reflection" section in this document for examples
223
228
- Other examples: datastore size-on-disk, ongoing performance and resource usage data
224
229
- Surface these in diagnostic panels, widgets, HUDs, etc
225
230
226
-
- [ ] API for logging outputs to datastore (time series data, feeds)
227
-
228
-
uses
229
-
- [ ] tag streaks -> atproto streaks (feeds + daytum)
231
+
Use-case: Tag streaks, group change feeds
232
+
- Pattern: aggregating sysetm user data
233
+
- tag streaks - eg one day you add tags "pushups, 10", next day "pushups, 20", then be able to query feed data for number of days in a row, max pushups, graph of pushups over time, longest streak, etc
234
+
- show in widgets
235
+
- push individual records or any rollup as an atproto record to get social streaking
230
236
231
237
## Entity centrism (NER streams)
232
238