Talk:Requests for comment/Performance standards for new features

Jump to navigation Jump to search
"There's no reason why we can't get to 1 s page load times" [median?, from 2-3 in last year]

Working on this[edit]

I am working on writing some performance guidelines and should have them ready to discuss and approve by the Zurich hackathon on May 9th. Sharihareswara (WMF) (talk) 13:25, 30 April 2014 (UTC)

Thanks. What of the models proposed in the RfC are you following for now? User:Sharihareswara (WMF)/Performance guidelines seems to reproduce the old usual focus on server-side performance i.e. "am I going to/did I break the cluster", is the goal to produce guidelines that would let people confidently merge code (especially by volunteers) without being blocked on a performance review by experts?
For the frontend performance side, are you going to have some more interviews/whatever or is that the material you're going to have for now? This part has been traditionally neglected and never has easily visible counter-proofs like "our database crashed", so it will need more compelling arguments and advice. --Nemo 14:11, 30 April 2014 (UTC)
Will look at this comment today. For more information please see Performance guidelines (which is a draft right now). Sharihareswara (WMF) (talk) 11:41, 9 May 2014 (UTC)
Hi. I don't think any of the thoughts mentioned in the current RfC are "models". The Performance guidelines I am working on constitute eight principles and will include details on how to achieve each of those principles, including good and bad examples. The goal of this document is to guide developers who are writing new code, and to guide code reviewers in becoming more expert in checking for performance impact. I do not plan on including several "must be this fast" metrics; I think that will require more thought and consensus by other people. That may require a separate document and thus my work may not be sufficient to close this RfC. As you can now see, I have expanded on the draft/notes that you found in my userspace, and am including material on frontend performance, backend performance, and persistence and caching layers. The recent attention to the front-end page load timing chart seems pretty compelling to me. Sharihareswara (WMF) (talk) 19:45, 9 May 2014 (UTC)
I've talked about this a bit with Ori and understand that Performance guidelines is probably separate from the standards that this RfC is about. So although Performance guidelines will probably be a useful document as context for the performance *standards*, I rescind my earlier implication that the drafting of Performance guidelines would fulfill/implement Requests for comment/Performance standards for new features. Sharihareswara (WMF) (talk) 23:46, 10 May 2014 (UTC)
Based on today's meeting, I think there is still some confusion whether that document should be (quoting) "aspirational rather than descriptive". --Nemo 08:55, 11 May 2014 (UTC)


"Know these things about your code: ...Reload the page. Does any interface element appear after the page has loaded? Does content shift or do interface elements move after DOMContentLoaded?"

Did you mean reflow instead of reload? Otherwise I do not understand this comment at all. --Tgr (WMF) (talk) 19:25, 8 May 2014 (UTC)