bigbones

Except for any media you post which end up as a unique key behind CloudFlare or Google Cloud load balancers. They may not reveal your chats, but they certainly export your metadata (i.e. list of devices fetching and putting a key) to the worst possible parties

Not a comment on Blitz itself, but had never heard of Dioxus before. Is there some unwritten rule about compiles-to-wasm type frameworks where they never actually show a demo or have their own site self-hosted in the framework? Seen this like 5 or 6 times now. I see some wasm file loaded on the Dioxus home page, but it's not clear where it's being used if at all

Because it would defeat the purpose. Turns out we don't trust the systems folks all that much

I didn't mean to imply avoiding use of a database entirely, almost any DB system tasked with copying a few long strings around in a simple query won't perform much worse than a literal raw disk file.

Even just something like: CREATE TABLE calendar(id, user_id, blob)

It is also a data model that amounts to a de facto standard across all real calendaring apps. The text serialization sure is ugly and who would want to touch it unless they had to, but that's just a parsing problem. The wiki page shows 14 component types, who is ever going to think of that many when designing a replacement schema? Why not just use the data model you already have 'for free'

A random event from my calendar serializes to 740 bytes of iCalendar. An extremely busy calendar containing one event every 15 minutes 9am-5pm 7 days per week only contains 11680 entries for a full year, which works out to around 8MB.

Would suggest instead of designing a schema at all, a calendar is a good example of a problem that might be far better implemented as a scan. Optimizing an iCalendar parser to traverse a range of dumped events at GB/sec-like throughputs would mean the above worst-case calendar could be scanned in single-digit milliseconds.

Since optimizing a parser is a much simpler problem to solve once than changing or adding to a bad data model after it has many users, and that the very first task involving your new data model is probably to write an iCalendar importer/exporter anyway, I think this would be a really great trade off.

> Meanwhile in AWS you would pay $27k a month to have the same IOPS as a Lenovo Thinkpad X1.

Curious about how this number was arrived at and what products were involved

Sounds like a source-unavailable version of Quickwit? https://quickwit.io/

Guessing it's a feature of the WebRTC stack if it's to be found anywhere, there's always a requirement for cancelling feedback from other meeting participants