5 ms·

So they built their own chat room feature? I feel that's a bit short-sighted, but I'll admit I'm only a random opinionated passer-by. You can certainly do it,

So they built their own chat room feature? I feel that's a bit short-sighted, but I'll admit I'm only a random opinionated passer-by. You can certainly do it, but you'll run quickly into issues and be buried in feature requests because like it or not, a chat room isn't just "a couple messages being passed back and forth" Hand rolling your own communication tool is a great way to become so busy you don't have time for your actual work.

barkingcat 6y ago
In the context of fossil-scm, it does make sense. The idea of fossil is that it's 1 embedded database (sqlite) that contains the entirety of your development activities. scm, wiki, defect tracking, a little bit of project management, and now chat room. The entire idea of fossil is that you shouldn't need to use another website or another tool for anything related to tracking your development. It's the "I am stranded on a tropical island, what do I use for hacking on a project if I can't have github or irc or slack or any other tools out there." In the context of fossil, chat is just a couple messages being passed around with no account taken at all for any other concerns. It's not a "chat" tool, it's a chat tool for the fossil-scm. Especially if you look at the Ephemeral feature - they delete chat messages after a certain configurable time - so they don't even have to worry about keeping messages.
mscrnt 6y ago
It's only for devs with commit bits—not random users. So not likely to be buried in feature requests and much more likely to be just a couple messages being passed back and forth between developers of the project. I feel it streamlines the process and improves cohesion a great deal; more of a long-sighted deal—but I'm only a random opinionated passer-by.
supportlocal4h 6y ago
Heh. Building an scm is a great way to become so busy you don't have time for your actual work. Building a db is another great way. If I could be so unproductive...
- sgbeal 6y ago
  > If I could be so unproductive... It goes much further than that... When one views a fossil-hosted forum post from the main fossil site or sqlite's site, they are looking at... - A forum post rendered by software Richard wrote. - Piped out to you via an HTTP server he wrote. - Served from an SCM he wrote. - Stored on a database package he wrote. - All coded in a text editor he wrote. Complete vertical integration. He's written, or had a majority hand in, every part of the chain except for the browser. We're all awaiting his announcement about his browser project. It's only a matter of time.
  - ephaeton 6y ago
    I'd rather he'd announce his MTA & MUA so we can reinstate the fossil mailing list. Going to its own quirky (at the time, haven't checked since the switch) forums meant I lost interest in being part of the (dev) community.
    - sgbeal 6y ago
      > I'd rather he'd announce his MTA & MUA so we can reinstate the fossil mailing list. It's demonstrably become impossible to manage spam in such an environment without the resources of mega-corps, whereas fossil aims to be a entirely SCM environment in a single binary hosted from modest hardware (e.g. a Raspberry Pi). The inability to manage spam was the driving factor for the forum. It is somewhat quirky, but it works well and our spam rate has been literally one single spam in nearly three years of operation (and that one was subtle - hiding an ad link inside a markdown hyperlink which enclosed only a single period). Similarly, the sqlite forum (which was migrated only relatively recently) is also spam-free. If you have found a way to develop a fully spam-proof, mailing-list-based solution then, by all means, show us the code. Until then, the forum has proven to completely solve the problem it set out to: provide a spam-free communication channel for fossil and sqlite.
  - gglitch 6y ago
    What's the text editor? If anything could get me off Emacs, it'd come from the Hippoverse.
sgbeal 6y ago
> I feel that's a bit short-sighted, ... You can certainly do it, but you'll run quickly into issues and be buried in feature requests (Disclaimer: i wrote 99%+ of the fossil chat front-end code.) FWIW... When Richard (our Alpha Coder) first posted about his proof-of-concept app for the chat feature my inner reaction was very similar to yours. However, writing a chat front-end sounded like an interesting way to spent Christmas, so dove right in. Now, exactly 3 months later, we've been using chat 24/7 within the Fossil project and i use it heavily on one of my own projects (hosted over CGI, and my biggest initial skepticism was whether chat could work reasonably well that way). Despite my early doubts it's been a genuine help to us within the project. i don't recall us getting more than 1 small feature request since then, and that was cut off the same way the hypothetical deluge would be: point them to the feature's scope document and redirect such users to Discord or Facebook or whatever the cool kids are using these days. It's been remarkably low-maintenance so far. The only notable development in the /chat code in the past 3 months was the recent replacement of window.fetch with more conventional XHR because of compatibility issues with window.fetch for some clients, and that took all of 90 minutes or so to do.
imtringued 6y ago
I had to build chat functionality into my own project. All I support is text and image/webm uploads. No, emojis. No edits, no quoting, no direct messaging. I wish there was a standard for this with implementations in every programming language. The only thing that I want to build myself is the UI.
- julianlam 6y ago
  Sure, but if there was a standard, then building the UI would be a massive undertaking, for the standard would be all-encompassing. Building a UI for a minimalist chat interface is easy, but if the standard supports all the fluff features that some people want...

barkingcat 6y ago

In the context of fossil-scm, it does make sense. The idea of fossil is that it's 1 embedded database (sqlite) that contains the entirety of your development activities. scm, wiki, defect tracking, a little bit of project management, and now chat room. The entire idea of fossil is that you shouldn't need to use another website or another tool for anything related to tracking your development. It's the "I am stranded on a tropical island, what do I use for hacking on a project if I can't have github or irc or slack or any other tools out there." In the context of fossil, chat is just a couple messages being passed around with no account taken at all for any other concerns. It's not a "chat" tool, it's a chat tool for the fossil-scm. Especially if you look at the Ephemeral feature - they delete chat messages after a certain configurable time - so they don't even have to worry about keeping messages.

mscrnt 6y ago

It's only for devs with commit bits—not random users. So not likely to be buried in feature requests and much more likely to be just a couple messages being passed back and forth between developers of the project. I feel it streamlines the process and improves cohesion a great deal; more of a long-sighted deal—but I'm only a random opinionated passer-by.

supportlocal4h 6y ago

Heh. Building an scm is a great way to become so busy you don't have time for your actual work. Building a db is another great way. If I could be so unproductive...

sgbeal 6y ago
> If I could be so unproductive... It goes much further than that... When one views a fossil-hosted forum post from the main fossil site or sqlite's site, they are looking at... - A forum post rendered by software Richard wrote. - Piped out to you via an HTTP server he wrote. - Served from an SCM he wrote. - Stored on a database package he wrote. - All coded in a text editor he wrote. Complete vertical integration. He's written, or had a majority hand in, every part of the chain except for the browser. We're all awaiting his announcement about his browser project. It's only a matter of time.
- ephaeton 6y ago
  I'd rather he'd announce his MTA & MUA so we can reinstate the fossil mailing list. Going to its own quirky (at the time, haven't checked since the switch) forums meant I lost interest in being part of the (dev) community.
  - sgbeal 6y ago
    > I'd rather he'd announce his MTA & MUA so we can reinstate the fossil mailing list. It's demonstrably become impossible to manage spam in such an environment without the resources of mega-corps, whereas fossil aims to be a entirely SCM environment in a single binary hosted from modest hardware (e.g. a Raspberry Pi). The inability to manage spam was the driving factor for the forum. It is somewhat quirky, but it works well and our spam rate has been literally one single spam in nearly three years of operation (and that one was subtle - hiding an ad link inside a markdown hyperlink which enclosed only a single period). Similarly, the sqlite forum (which was migrated only relatively recently) is also spam-free. If you have found a way to develop a fully spam-proof, mailing-list-based solution then, by all means, show us the code. Until then, the forum has proven to completely solve the problem it set out to: provide a spam-free communication channel for fossil and sqlite.
- gglitch 6y ago
  What's the text editor? If anything could get me off Emacs, it'd come from the Hippoverse.

sgbeal 6y ago

> I feel that's a bit short-sighted, ... You can certainly do it, but you'll run quickly into issues and be buried in feature requests (Disclaimer: i wrote 99%+ of the fossil chat front-end code.) FWIW... When Richard (our Alpha Coder) first posted about his proof-of-concept app for the chat feature my inner reaction was very similar to yours. However, writing a chat front-end sounded like an interesting way to spent Christmas, so dove right in. Now, exactly 3 months later, we've been using chat 24/7 within the Fossil project and i use it heavily on one of my own projects (hosted over CGI, and my biggest initial skepticism was whether chat could work reasonably well that way). Despite my early doubts it's been a genuine help to us within the project. i don't recall us getting more than 1 small feature request since then, and that was cut off the same way the hypothetical deluge would be: point them to the feature's scope document and redirect such users to Discord or Facebook or whatever the cool kids are using these days. It's been remarkably low-maintenance so far. The only notable development in the /chat code in the past 3 months was the recent replacement of window.fetch with more conventional XHR because of compatibility issues with window.fetch for some clients, and that took all of 90 minutes or so to do.

imtringued 6y ago

I had to build chat functionality into my own project. All I support is text and image/webm uploads. No, emojis. No edits, no quoting, no direct messaging. I wish there was a standard for this with implementations in every programming language. The only thing that I want to build myself is the UI.

julianlam 6y ago
Sure, but if there was a standard, then building the UI would be a massive undertaking, for the standard would be all-encompassing. Building a UI for a minimalist chat interface is easy, but if the standard supports all the fluff features that some people want...