37

It's a feature request almost as old as the network itself - here's a request for a streaming API that felt like it got off on the wrong foot, and a frustratingly cryptic answer from Jeff when asked if chat would be included in the API

It's a testament both to the robustness of the chat system, and of the skills of folks who reverse engineered various libraries underpinning useful tools like SmokeDetector that chat's worked reliably over all this time, with tools rarely breaking.

Amusingly while back in the day, one of the arguments for not having a front end API was "we might break something if we change it" - changing the front end has and will break things. "it seems to formally legitimise the scripts" - which essentially seems to be overdue with things like SmokeDetector.. and so on. That "the chat system is evolving" as said in the Stack Apps post isn't particularly true either, with no major changes in chat other than small changes for Teacher's Lounge moderations, and before that the 'new' mobile view.

It's not a state of affairs I'd consider ideal.

We have critical tools to the community running on the basis that there's no changes to the chat system, running on screen scraping or JavaScript witchcraft. Changes might break things, which would reduce the effectiveness of tools like SmokeDetector/Charcoal and various other utility bots used by mods and others. Basically, folks have done their best with what we have, but... this is somewhere where Stack can do better.

While I realise some 'internal' SE tooling has moved on to other (non inhouse/ inferior?) chat systems which have their own chatops tooling, there's probably multiple benefits to a documented, robust API that's not put together from reverse engineering and careful jerryrigging both for internal SE use and for community projects.

I suspect for chat to be a 'going concern' - there's going to be a point where it would probably need to be worked on, bringing it in line with the regular Q&A sites. I do realise this isn't on the current roadmap, but things like moving the front end to stacks might ease maintenance in the future, not to mention not leaving the code base to moulder, unloved until something catastrophic breaks, or there's a critical feature addition.

This might, of course, break many bots, and in preference to keeping things 'status quo' during a longer transition, giving folks the tools to move away from the current reverse engineered pseudo APIs/libraries to 'proper' APIs would be a very positive step.

While I realise there's (often?) other priorities, I do hope that updates to chat, quality of life improvements as well as planning for the future is something that might be considered.

A chat API would make a lot of these easier and hopefully help form a solid foundation for future of chat.

17
  • Does this answer your question? Will chat be accessible through the API? Nov 18, 2021 at 14:28
  • 4
    @EkadhSingh-ReinstateMonica why would it answer the question about reconsidering exactly this thing?
    – VLAZ
    Nov 18, 2021 at 14:32
  • @VLAZ they’re still asking for exactly the same thing. Nov 18, 2021 at 14:34
  • 12
    @EkadhSingh-ReinstateMonica : 11 years ago, there was no smoke detector and a load of other tools that use chat as a front end. Chat was new and shiny. We didn't have any libraries or tools that relied on the front end not changing to work effectively. And frankly - while it was ok for the time - Jeff's answer wouldn't really pass muster in these days. There's a tangential/related item on the 2021 roadmap as well, so when it gets started, It would be good to have an up to date post asking for this Nov 18, 2021 at 14:50
  • 3
    So, I kinda feel that not only is it not a duplicate - considering what I'm trying to do, closing the duplicate in that direction both closes a more detailed post for a less detailed one, and removes an avenue to bring up the issue for reconsideration considering the responses in the past. I'd also add, and I have a bias here, faced with a 'better' newer question, and a less good old one, its ok to duplicate vote the other way and close the older question against the newer one. Nov 18, 2021 at 14:53
  • I have the first linked... the second seems like something I missed, and can be rolled into my question Nov 18, 2021 at 15:09
  • 2
    @EkadhSingh-ReinstateMonica No, they are not. The question here clearly acknowledges the previous one and points out how it's different to it due to different circumstances. Moreover, the company and product has changed. Jeff is not even on a leadership position here any more. Unless you want to appeal to tradition "it has always been this way therefore, it must never change", I do not see a reason to close this Q as a dupe to the other one. I should also mention that some things did change from the way they were, so that wouldn't even be a good argument against the reconsideration.
    – VLAZ
    Nov 18, 2021 at 15:38
  • My vote: close the old question as a duplicate of this one.
    – Laurel
    Nov 18, 2021 at 23:03
  • Any of these useful/critical tools matter for Q&A? (Not a rhetorical question.) Nov 19, 2021 at 20:27
  • 2
    @Scortchi-ReinstateMonica would you consider preventing spam like the current aggressive campaign going on right now "useful" for the Q&A? Or all the other efforts to keep the site clean like SOBotics? I suppose they aren't critical, nor are they something that directly relates to Q&A but I think they are useful to the site.
    – VLAZ
    Nov 19, 2021 at 20:34
  • (1) Yes; (2) I don't know what SOBotics is. I've heard of SmokeDetector/Charcoal, but know only that they're to do with preventing spam & was unaware they somehow depend on chat to work. That might be shockingly ignorant, but I doubt I'm alone, & it might be worth explaining why this proposal should matter to people who don't use or care about chat. Nov 19, 2021 at 22:08
  • I didn't want to get into the weeds of "these exactly" are the things affected, probably cause I think that chat is a integral part of the stack exchange network. A few examples though might be realtime feeds of suspicious posts that smoke detector/charcoal does, for folks to review and flag accordingly. I know of one site that (used to?) have a bot that scanned new comments and brought up potentially toxic ones for review. SOBotics hosts a bunch of libraries for folks to use for tools like that. Very fundamentally, I think work needs to be done on chat for it to be sustainable in future Nov 20, 2021 at 3:10
  • 2
    and not be a giant pile of tech debt that's basically shambling on because 'it runs' until it doesn't. Nov 20, 2021 at 3:11
  • 2
    @Scortchi-ReinstateMonica A huge number of volunteer efforts to support/clean the entire network up exist and are managed almost solely from chat. Improving the ability to work with the site via chat would greatly improve the ability of these volunteer efforts to keep network sites clean of spam, abuse, etc.
    – TylerH
    Dec 16, 2021 at 16:37

0

You must log in to answer this question.

Browse other questions tagged .