Gwern Branwen wrote:
On 2007.10.08 12:17:52 -0700, Ray Saintonge scribbled 70 lines:
Gwern Branwen wrote
I don't think there's really anyway to solve this. Nobody is really advocating putting that kind of metadata into the article, which would be a herculean and sisyphean task; nobody is seriously talking about associating a second page with articles (one for discussion and the other for metadata)
It does little good to belittle the idea of a separate metadata page as not serious when you don't even give arguemnts about why it's such a bad idea.
It's not necessarily a bad idea, it's just that if we're going to discuss changes to ameliorate this situation, let's focus on ones which actually have a chance of getting implement. It's a fact of life here at En that large-scale technical changes, however awesome they may be, usually don't happen and if they do it takes *forever* (because of many factors ranging from inertia to not wanting to clean up the ensuing mess to various developer issues to cruft, etc. etc.). This is true even for really really important things like SUl, or - actually, this might be going a little bit too far back here, but does anyone else remember Bug 550? That was an incredibly annoying limitation which took years to change.
Anyway, the essential point is that user-run bots and scripts got us into this mess and they seem to represent the most realistic way out.
That's always a major limitation. Bug 550 doesn't ring a bell, but it would be easy for a Bug number to pass me by. The SUI problem remains there, as does the development of an internal search function that doesn't suck, and as do others. We certainly do not lack technical talent. It's also important that maintenance issues have priority on the developers time. It appears that it's the decision making process that is in paralysis; that makes it easier to use short-sighted policy workarounds.
But I think in a certain modest way there is a solution. Just display the talk automatically. At the bottom is a good place. Think about it: if it's some worthless banners and templates, you simply stop reading/scrolling-down at the categories - but if there is a lot left, then you continue reading and merely skip over the templates. The additional load time is negligible, it doesn't mess anything up, etc. And it's a relatively simple addition to one's monobook.js: https://secure.wikimedia.org/wikipedia/en/wiki/Wikipedia:WikiProject_User_scripts/Requests#Automatically_view_talk_page_discussion_while_looking_at_article. I'm not suggesting it be put into the site-wide file, but I think it could be a step towards a programmatic solution for annoyed editors.
Not everybody wants to see the talk page, or even to load such a long page. We already try to keep the article sizes down for the benefit of people with slow browsers. An active talk page can be very, very long, and can even have many archive pages. This idea is not much different from sticking the metadata on the article page itself.
Well, I think browsers would usually begin rendering the frame containing the talk page after the article proper.
.Some old browsers tend to be unresponsive even to such basic actions as scrolling until the whole file is loaded.
If the idea had any sort of usefulness it would need to be opt-in, because we can't expect everybody to be able to make sense of monobook.js. The ones who really need to use the talk page are not the high-tech types that understand monobook.js; they are those editors who may understand and are familiar with content while being mystified with anything more complicated than the most basic wiki syntax.
I'm not saying it shouldn't be opt-in - 'I'm not suggesting it be put into the site-wide file...' - but I think a more advanced version of this (or at least *something* that would either recolor the talk link when it's all bots, or transclude it if bots, or... really there are a lot of solutions here) would be good to have generally available to the community.
Intuitively, something like this seems much easier for segregating bots and templates, or we already create archives for long talk pages. Could a metapage be treated as a kind of special archive.
For example, it's an old and perennial proposal to give rollback to non-admin editors, but it never has flown and never will out of sheer inertia even if consensus is for it - but the people who need rollback while not an administrator now have nice opt-in scripts like the TWINKLE stuff which are almost as good as real rollback.
Inertia remains a human problem. Technical solutions can be subject to the same kind of ownership and protection problems as the articles themselves. Most of us can have a basic (if erroneous) understanding of the issues in a political debate, but that is not the case with technical issues.
Ec