|
2017-11-22 02:43:32
|
<stjn>
|
I think I should open an RfC on software bloating the Wikimedia infrastructure on this point -_-
|
|
2017-11-22 02:44:29
|
<stjn>
|
Given that how many actions now are being taken by ‘enough users haven’t objected in a fast time and we listened only to those that gave us positive feedback’.
|
|
2017-11-22 02:45:47
|
<Isarra>
|
Hi stjn! Is this about timeless?
|
|
2017-11-22 02:46:11
|
<stjn>
|
Not only, the RevisionLoader was also deployed the same way for example.
|
|
2017-11-22 02:47:48
|
<stjn>
|
But yes, I think that asking users first about ‘testing stuff’ and then deploying it as a default feature everywhere despite them giving a permission only to test is a bad way of conducting development.
|
|
2017-11-22 02:47:59
|
<Isarra>
|
Both of these are things you can just turn off (and in timeless' case, isn't even on by default).
|
|
2017-11-22 02:48:47
|
<stjn>
|
It does not mean that Wikimedia infrastructure should contain hundreds upon hundreds of extensions, skins etc. no matter how useful or polished they are.
|
|
2017-11-22 02:48:53
|
<Isarra>
|
Depends on the development. If it's a generall improvement, why wouldn't it be turned on by default?
|
|
2017-11-22 02:49:19
|
<Isarra>
|
Are you a developer?
|
|
2017-11-22 02:50:40
|
<stjn>
|
Not in PHP, so I guess I can’t go on and stick anything that would stick to MediaWiki :-D
|
|
2017-11-22 02:51:03
|
<Isarra>
|
Well, I'm wondering more in general.
|
|
2017-11-22 02:51:38
|
<legoktm>
|
[18:48:43] <stjn> It does not mean that Wikimedia infrastructure should contain hundreds upon hundreds of extensions, skins etc. no matter how useful or polished they are. <-- if they're useful, why not?
|
|
2017-11-22 02:51:41
|
<stjn>
|
I really don’t want to undermine your or WMDE’s efforts, but generally this fast-tracking got so bad I don’t know how to put up with it anymore.
|
|
2017-11-22 02:51:56
|
<Isarra>
|
Because it's pretty common to try to break things up into separate modules and the like so they don't wind up all relying on each other. Having all these different extensions and skins lets us experiment and try to improve the overall perfomance and user experience of the sites without them winding up too dependent on each other.
|
|
2017-11-22 02:52:26
|
<Isarra>
|
So there being a lot of them is generally a good thing, as opposed to it all being random junk in core, or the like...
|
|
2017-11-22 02:53:02
|
<legoktm>
|
how fast is too fast? the initial request to deploy timeless was filed on Dec 31, 2016. nearly 11 months ago.
|
|
2017-11-22 02:56:02
|
<stjn>
|
legoktm: because site should be simpler to use, not harder. Every feature that is being stuck here and there does increase the common weight, and I am not talking about the skin.
|
|
2017-11-22 02:56:03
|
<bawolff>
|
As far as software bloating goes, I don't really think one option skin and one js thingy that's entirely independent of everything else is really very much
|
|
2017-11-22 02:56:40
|
<bawolff>
|
I'd be much more conerned about highly coupled components
|
|
2017-11-22 02:56:42
|
<stjn>
|
About the skin, more generally, I have never heard any comments about whether or not design review would be even taken in the consideration. At no point, as far as I recall.
|
|
2017-11-22 02:57:39
|
<legoktm>
|
stjn: agreed, but a responsive skin seems like it would make the site easier to use - one UI regardless of platform. Plus it's forcing us to write skin-agnostic code which means less coupling and tech debt
|
|
2017-11-22 02:57:44
|
<Isarra>
|
stjn: You mean besides my several comments where I was specifically following up on it, saying I would address specifics, filing bugs for the points raised?
|
|
2017-11-22 02:58:06
|
<bawolff>
|
Well there was a design review at https://phabricator.wikimedia.org/T158012 and it seems to have been considered
|
|
2017-11-22 02:59:04
|
<Isarra>
|
I haven't actually gotten around to working on it, but that's a lot of work. Which is kind of why I was asking for funding so that I could prioritise doing that...
|
|
2017-11-22 02:59:16
|
<Isarra>
|
Because it kind of needs it...
|
|
2017-11-22 03:00:20
|
<stjn>
|
legoktm: again, I am not objecting at the point of developing this skin or a responsive skin in general (it’s a marvelous effort and can be only respected), I am objecting against releasing it in the wild despite the fact that it is simply not polished enough yet. I mean, I could probably even use this skin at some point in the future, it doesn’t mean that the projects that didn’t request it specifically should get it now because there
|
|
2017-11-22 03:01:02
|
<bawolff>
|
Its only available by selecting something on your Special:Preferences
|
|
2017-11-22 03:01:11
|
<bawolff>
|
Its not like we turned it on by default
|
|
2017-11-22 03:01:18
|
<legoktm>
|
^
|
|
2017-11-22 03:01:35
|
<bawolff>
|
I agree its not perfect, but I think its good enough
|
|
2017-11-22 03:01:46
|
<stjn>
|
Yes, but local developers like me and others in that conversation are expected to support any current skins at any point, even if they are not default.
|
|
2017-11-22 03:02:07
|
<Isarra>
|
And you can't effectively develop this stuff in a vacuum. It doesn't work. You HAVE to release it in order to actually see how users interact with it, what bugs occur with everything else deployed, etc, in order to even know WHAT to work on and continue developing.
|
|
2017-11-22 03:02:28
|
<Isarra>
|
What do you do that needs to support these skins?
|
|
2017-11-22 03:03:02
|
<stjn>
|
I am not talking about the vacuum (maybe in some comments that was the impression, I am sorry for that), those projects that have requested it are very diverse and not a vacuum-like already, no?
|
|
2017-11-22 03:05:10
|
<legoktm>
|
stjn: Erm, who is expecting you to support any current skins at any point?
|
|
2017-11-22 03:05:44
|
<Isarra>
|
As previously stated, with more and more projects asking to have it enabled faster than we can deploy it, it made more sense to simply enable it everywhere instead and let anyone who doesn't want it opt out.
|
|
2017-11-22 03:06:26
|
<stjn>
|
Users in local communities? I mean, if tech admins at RuWP simply say ‘your loss, we don’t support your %non-Vector-or-Monobook%’, how exactly are you expecting it to go down?
|
|
2017-11-22 03:07:26
|
<bawolff>
|
If the users don't like it, you tell them tough, and if they want they can {{sofixit}}
|
|
2017-11-22 03:08:13
|
<Isarra>
|
That's... fine? Not all gadgets work in all skins. That's pretty much always been the case as long as there have been gadgets.
|
|
2017-11-22 03:08:46
|
<Isarra>
|
Are you a gadget developer, then?
|
|
2017-11-22 03:10:33
|
<stjn>
|
Sorry if I am taking too long to answer. Bug with Modern was discovered, for example, after we moved coordinates into indicators and started to use Wikimedia maps, so maplinks got illegible from the background colour there. And problems here and there are frequent and these kinds of decisions increase the burden without anything in return (given that the problems in the skin on current point prevent using it in a long-term).
|
|
2017-11-22 03:11:34
|
<stjn>
|
Isarra: I am one of technical admins (‘engineers’ in local translation) at Russian Wikipedia, yes (so I have access to both MediaWiki configuration and protected templates/modules).
|
|
2017-11-22 03:14:00
|
<Isarra>
|
Has the modern bug been filed on phabricator?
|
|
2017-11-22 03:15:07
|
<Isarra>
|
Okay, so you're concerned about this creating more work for you for the template styles?
|
|
2017-11-22 03:15:32
|
<stjn>
|
WMF by then already soon-to-be-closed Maps department, so no.
|
|
2017-11-22 03:17:15
|
<bawolff>
|
Not soon to be closed
|
|
2017-11-22 03:17:21
|
<bawolff>
|
It was closed like 6 months ago
|
|
2017-11-22 03:17:55
|
<stjn>
|
*By then*.
|
|
2017-11-22 03:18:52
|
<Isarra>
|
I'm not really following. Is the bug no longer present, then?
|
|
2017-11-22 03:21:10
|
<stjn>
|
‘Okay, so you're concerned about this creating more work for you for the template styles?’ − just work with any styles, given that even the search bar is not in the right place for me and there are too many other bugs that would have to be filed or fixed locally by someone.
|
|
2017-11-22 03:21:54
|
<Isarra>
|
If you're not even going to file the bugs, why the hell are you complaining about it making you work? Just ignore the damn thing and be done with it.
|
|
2017-11-22 03:23:20
|
<stjn>
|
‘Is the bug no longer present, then?’ − well, we have a local override for it so that the text at least is legible.
|
|
2017-11-22 03:24:27
|
<ashley>
|
as with all bugs, developers -- no matter how good they are -- are not mind readers, so if you don't inform them of the bugs, they won't be able to fix it as they're not aware of the bugs until you or someone else makes 'em aware
|
|
2017-11-22 03:24:28
|
<bawolff>
|
Look, here's the way it is - If you don't file bugs but fix them with js hacks, we are not going to fix them because we don't know about them
|
|
2017-11-22 03:24:51
|
<bawolff>
|
If you don't want it on ru projects, we'll take it off, if there is consensus on that wiki
|
|
2017-11-22 03:26:33
|
<ashley>
|
wanting to disable an *opt-in*, non-default feature that is not actively harming anyone but rather allows users to try the feature out and provide feedback on how to improve it seems rather strange to me
|
|
2017-11-22 03:31:55
|
<stjn>
|
CSS hacks, mind you. Other than that, I really see no point in this conversation anymore − it’s clear that individual opinions don’t matter anymore to the point that developers yet again have the audacity to tell me to seek community consensus when they are somehow not bound by it.
|
|
2017-11-22 03:33:25
|
<bawolff>
|
"developers yet again have the audacity to tell me to seek community consensus when they are somehow not bound by it." <-- Yep that pretty much describes the situation
|
|
2017-11-22 03:34:24
|
<ashley>
|
it's rather obvious that there cannot be a consensus on every change ever, or otherwise we'd still be working on MediaWiki 1.1 and debating over whether it should support extensions or not. as bawolff said, community consensus is the key but I'm not sure why you're eager to axe an *opt-in* feature instead of trying to work to improve it. the wonder of collaborative open source projects is that software is made better by people working
|
|
2017-11-22 03:34:24
|
<ashley>
|
together for a common cause
|
|
2017-11-22 03:36:47
|
<Isarra>
|
You're one person arguing with a bunch of people. Get a bunch of people as well and show it's not just you, and we'll probably respect that, but as is, you're right - continuing to argue this isn't going to change anything.
|
|
2017-11-22 03:36:53
|
<Isarra>
|
But that's sort of how consensus works.
|
|
2017-11-22 03:37:22
|
<stjn>
|
I would’ve been keen to testing this and maybe reporting some if there was some indication that this is not a final state of the project and that this is not the state in which it will be presented to the general public across the Wikimedia movement. But I never got that, the idea always was always to stomp through to the point of a global option.
|
|
2017-11-22 03:39:51
|
<stjn>
|
I mean, I am not in charge of the development process and I am not defending my decisions, so a rather Anglocentric point of view. I could’ve got the same kind of consensus among our tech team that Timeless should not be enabled; alas, they are not the whole community and neither are you.
|
|
2017-11-22 03:41:13
|
<Isarra>
|
It's not the final state of the project. It's not even close. Where did you get that idea?
|
|
2017-11-22 03:43:57
|
<Isarra>
|
What we've deployed is basically the result of 15 hours of me making a tutorial skin, plus about two year's worth of bugfixes, but there's a damn reason why I'm asking for funding now to properly work on it. We're getting good feedback and there's a lot to work on, and now we even have an idea what people actually do want from it.
|
|
2017-11-22 03:44:04
|
<Isarra>
|
If anything, we're just *starting*.
|
|
2017-11-22 03:44:51
|
<bawolff>
|
stjn: Which is precisely why we're not removing timeless just based on what you're saying. We're willing to remove it if ru.wikipedia.org doesn't want it. We don't care if they russian wikipedia "tech team" (whatever that is) doesn't want it
|
|
2017-11-22 03:45:26
|
<stjn>
|
Well, it is rapidly approaching success according to some (despite a very lacking development). So, not to argue anymore about something I can’t change until I could influence inner circle decisions: would this be announced in tech news?
|
|
2017-11-22 03:46:13
|
<stjn>
|
bawolff: Relax, it was just an analogy to your decision making process that got us here in the first place.
|
|
2017-11-22 03:54:08
|
<bawolff>
|
stjn: Its already been included in next week's tech news
|
|
2017-11-22 03:55:38
|
<bawolff>
|
Its not really my decision making process so much as following procedures that have been in place for over a decade
|
|
2017-11-22 15:03:22
|
<Nudin>
|
Technical Advice IRC meeting starting in 60 minutes in channel #wikimedia-tech, hosts: @Lucas_WMDE & @Thiemo_WMDE - all questions welcome, more infos: https://www.mediawiki.org/wiki/Technical_Advice_IRC_Meeting
|
|
2017-11-22 15:59:54
|
<Thiemo_WMDE>
|
Counting the seconds to start this weeks Technical Advice IRC Meeting... ;-)
|
|
2017-11-22 16:01:15
|
<Nemo_bis>
|
:)
|
|
2017-11-22 16:02:09
|
<Lucas_WMDE>
|
welcome everyone to another Technical Advice IRC meeting! your hosts today are: Thiemo_WMDE, Lucas_WMDE, and MichaelSchoenitz :)
|
|
2017-11-22 16:02:49
|
<zuzak>
|
duly applauds
|
|
2017-11-22 16:02:50
|
<Lucas_WMDE>
|
I’m not aware of any pre-requested topics, so just ask away
|
|
2017-11-22 16:04:52
|
<Thiemo_WMDE>
|
We will not present anything without being asked as this is not the idea of the following hour. ;-)
|
|
2017-11-22 16:07:09
|
<Aguyintobooks>
|
commons structured data --> will it definitely include EXIF data support
|
|
2017-11-22 16:07:37
|
<Thiemo_WMDE>
|
It will definitely include EXIF in a way that you can easily search for it.
|
|
2017-11-22 16:08:03
|
<Aguyintobooks>
|
brilliant
|
|
2017-11-22 16:08:55
|
<Thiemo_WMDE>
|
It's not decided how exactly EXIF data will interact with data from MediaInfo statements. For example, coordinates in EXIF tend to be wrong and there should be a way to fix them, but without messing with the original file.
|
|
2017-11-22 16:10:33
|
<Aguyintobooks>
|
also will there be a transition period where the new version of commons will be ported over from the old version, or will it be taken down for a limited period, or will you try and update the existing database in situ?
|
|
2017-11-22 16:10:37
|
<Nemo_bis>
|
Isn't EXIF already indexed by CirrusSearch
|
|
2017-11-22 16:11:27
|
<Thiemo_WMDE>
|
Nemo_bis: I believe so, but who knows how to search for it?
|
|
2017-11-22 16:12:40
|
<Thiemo_WMDE>
|
Aguyintobooks: There is not really an "old" and a "new" version. The plan is to add new functionality that *can* (if the community wants) replace old functionality from the {{Information}} template and category system. But the migration is mostly up to the community.
|
|
2017-11-22 16:13:16
|
<Lucas_WMDE>
|
there’s also not just one date where all the new functionality is available at once
|
|
2017-11-22 16:13:25
|
<Lucas_WMDE>
|
it’ll come in batches, some smaller features first
|
|
2017-11-22 16:13:26
|
<Thiemo_WMDE>
|
Aguyintobooks: What do you mean when you say "old database"?
|
|
2017-11-22 16:15:35
|
<Thiemo_WMDE>
|
Nemo_bis: Doesn't look like EXIF is currently indexed: https://phabricator.wikimedia.org/T157055 . However, it is in a secondary database table that can be queried via api.php.
|
|
2017-11-22 16:19:55
|
<Aguyintobooks>
|
well I was rather assumeing that you (they) would build a fully operational, but empty wikibase system, and simply use a bot to copy the files into it
|
|
2017-11-22 16:19:59
|
<Lucas_WMDE>
|
Aguyintobooks: for example, the first deployed feature will likely be captions, “a short, translatable line of text that explains what is in an image” (quoting spinster from yesterday’s Structured Commons office hour)
|
|
2017-11-22 16:20:57
|
<Thiemo_WMDE>
|
WMF will most probably not run such a bot. The community will.
|
|
2017-11-22 16:21:07
|
<Aguyintobooks>
|
yes but part of the concerns I have seen expressed are 1. its taking too long, and 2. all the features will have to be enabled semi-manually
|
|
2017-11-22 16:21:26
|
<Thiemo_WMDE>
|
What can take "too long"?
|
|
2017-11-22 16:23:03
|
<Aguyintobooks>
|
"Development takes place in 2017-2019." - many people, even me, would like to see at least a working mock-up sometime in 2018
|
|
2017-11-22 16:23:03
|
<Thiemo_WMDE>
|
Also, what should be the source? Thats why I asked what you mean. There is just no source a structured database can be fed from. Data must be extracted first, and that is usually done via bots.
|
|
2017-11-22 16:24:09
|
<Aguyintobooks>
|
which reverts to my original query, of having parallel databases
|
|
2017-11-22 16:24:18
|
<Thiemo_WMDE>
|
A working mock up is there for a while: http://structured-commons.wmflabs.org/wiki/MediaInfo:M13
|
|
2017-11-22 16:24:19
|
<stashbot>
|
M13: Cite Inspector - https://phabricator.wikimedia.org/M13
|
|
2017-11-22 16:25:49
|
<Aguyintobooks>
|
so at this point someone should put a link to that prominently on the homepage of the project
|
|
2017-11-22 16:25:52
|
<Thiemo_WMDE>
|
I don't understand what you want to have in "parallel". There is no old database. Only a new one, based on the Wikidata technology. Data for this database must be manually entered (e.g. what an image depicts, such thing is not stored anywhere). Other data must be extracted from templates. That is a manual process with a high margin of errors. These must be fixed manually.
|
|
2017-11-22 16:28:42
|
<Aguyintobooks>
|
well my concerns are primarily with the estimated time this is going to take, for example, you have a mock-up, how much effort would it take to make this fully operational, or is the 2 year development phase infact based on scaling it into Commons
|
|
2017-11-22 16:29:39
|
<Aguyintobooks>
|
possibly a concern of minimizing disruption?
|
|
2017-11-22 16:29:46
|
<Thiemo_WMDE>
|
I can not answer this because I'm not part of the WMF team that works on this right now.
|
|
2017-11-22 16:30:04
|
<Thiemo_WMDE>
|
I don't understand what you want to imply or what you are asking for.
|
|
2017-11-22 16:30:31
|
<Thiemo_WMDE>
|
Personally, I believe slower is *better*. Going slower generally reduces disruption.
|
|
2017-11-22 16:30:59
|
<Aguyintobooks>
|
yes it is probably sensible
|
|
2017-11-22 16:32:21
|
<Thiemo_WMDE>
|
I recommend to join the IRC office hour the WMF team does for this project. https://commons.wikimedia.org/wiki/Commons:Structured_data
|
|
2017-11-22 16:33:00
|
<Aguyintobooks>
|
yes good idea
|
|
2017-11-22 16:33:02
|
<Aguyintobooks>
|
thanks
|
|
2017-11-22 16:33:23
|
<Thiemo_WMDE>
|
The Wikidata team in Berlin (which I'm part of) started this project and still guides it, but it is a WMF project now.
|
|
2017-11-22 16:35:33
|
<Aguyintobooks>
|
yes there was major confusion on that the other day, until legotkm explained that it was a joint project
|
|
2017-11-22 16:36:45
|
<Thiemo_WMDE>
|
Basically, WMDE created the first prototype I showed you.
|
|
2017-11-22 16:38:09
|
<Thiemo_WMDE>
|
Other people, please feel free to interrupt us any time. We did not planned to exclusively talk about Commons, so if you have other topics, please go ahead.
|
|
2017-11-22 16:59:00
|
<Lucas_WMDE>
|
last call for any technical advice questions? ;)
|
|
2017-11-22 18:53:35
|
<krd>
|
Hello. Can anybody please sent me a traceroute from the Amsterday site (to a destination I'd provide in private)?
|
|
2017-11-22 18:53:54
|
<krd>
|
s/day/dam/ :-)
|