Want to take part in these discussions? Sign in if you have an account, or apply for one below
Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.
The rise of MathOverflow brings to the fore an important issue. We tend to think of the n-lab from the inside and how best to make it work for us. However, it is public, and it's being public is important and makes it more useful for us. So we should think a little about how we want "Joe Public" to interact with the n-lab.
Some random thoughts to start off:
The links across the top of the page are mostly useless for a casual visitor. Certainly "Recently Revised", "Authors", "All Pages", and "Export". Perhaps these should be at the bottom with the more author-focussed links (like "Edit" and "History") and at the top should be prominent links like "About" and "HowTo" and "FAQ". It's worth remembering that people don't always start their visit at the home page. We could include the main contents on every page, but we'd have to remember to do that each time whereas editing the main navigation bars is a once-for-all solution.
When someone links to the n-lab, what should happen? Can they link to any page, including pages in the history or "meta" pages like "recently revised"? If so, this is potentially dangerous since those pages are protected from search engines within the site but an external link is not so protected. Thus someone just needs to link to, say, the history of our infamous spam page and it instantly becomes indexable by search engines. Thus even if we blank or edit out links, their presence in our history becomes of some value. It would be possible to put a redirect in place that ensures that all external links always go to the current version of the page. The disadvantage of this is that I can see one situation where linking to history is useful: quoting in a paper. Then I may wish to say "see nlab page X (sourced: 18th September 2009)" to indicate that when I read that page then it did truly say what I am quoting it as saying even though it may have changed since. However, I can still refer to a particular revision even if I can't link directly to it so I don't see this as a major difficulty.
I'd like to have a short version of the "about" page (I realise that the current About is not up to date!) so that anyone wishing to put a description of the n-lab somewhere can simply cut-and-paste it and not have to make something up themselves.
If we (i.e.: you) can change the arrangement of the main navigation bars, there'd be good reason to do so, the way you indicate.
And, yes, it would be good to have an "about"-statement suitable for copy-and-pasting.
those pages are protected from search engines within the site but an external link is not so protected
Why not? One can have robots.txt tell them not to index pages just as much as one can have it tell them not to follow links. How can we fix that?
It is quite convenient to link to diff revisions when discussing them, perhaps from the Forum.
I wholeheartedly agree with making the top navigation bar more useful to new visitors.
There should also be a link to Latest Changes on every page.
How about a message on 'history' pages that makes it clear that a history page is a history page? Something like:
Warning: you are currently viewing an old version of this page. Click here for the latest version.
Secondly, I've been thinking a bit about navigation bars. It seems to me that there should be four groups of links: accessible, user, author, and page.
Accessible: in the normal run of things, these should be invisible. When viewing the page without a stylesheet (such as in a screen reader), these will come up first and so should be the "links you should be able to find without wading through the whole page". Obvious ones: home page and the about page, but perhaps some others as well. (NB CSS handles the invisibility very well.)
User: Links that a casual browser are likely to use. Suggestions: Home page, about, FAQ, HowTo, Search, Print view
Author: Links that an author might use. Suggestions: if FAQ and HowTo aren't on the User list then they should be here (duplication is not necessarily a bad thing), Recently Revised, Latest Changes, Authors (not sure about this one), Feeds, Source view (not yet implemented)
Page: Links that are specific to that page. Suggestions: Edit page, list of links from and includes.
All of these lists would be styleable so, by judicious use of CSS and either the Stylish Firefox or a more ambitious theme extension to Instiki, we could ensure that these menus are not obtrusive and do not get in the way of "ordinary use" (whatever that means).
The Wikipedias have been dealing with this for some time, so we can look at their solution for ideas. Check it out http://en.wikipedia.org/wiki/Higher_category_theory in lynx to see how they handle accessibility.
We should some day have a way to completely destroy spam pages or those versions which we all agree as unwanted. Then the problem of linking to these versions is not a problem any more.
We can delete pages and this completely removes them from the database. So far, no-one has particularly pushed for this option even on spam pages. I have the impression that Toby in particular is keen for us not to actually delete anything but just put it "beyond use". I would vote for deleting a page if it contained something offensive that we would not want anyone to encounter on the n-lab pages no matter how much digging they did.
I'm not absolutely against deleting things, but I want to avoid a situation where something borderline gets deleted without everyone realising what's going on. For example, if we decide that spam can be deleted and people start deleting spam as a matter of course, what happens when somebody decides that a page by a new JA is spam and deletes it, despite its (however tenuous and uninteresting) legitimate mathematical content? Then what happens if it's not someone like JA, but someone that we really want but the deleter did not recognise the content as interesting and relevant?
As long as it's done on an individual basis, I'd be willing to delete individual pages, say by consenus of the steering committee. On the other hand, if putting things "beyond use", as Andrew put it, does the job (and is something that any user can do), then why bother the steering committee?
What about the notification on history pages? I think that suggestion got lost in the rest.
On the other hand, if putting things "beyond use", as Andrew put it, does the jobž
I do not think that it does the job that completelym though it does to most purposes. For example, I already feel uneasy to advertise that lab has over 2500 pages, if I know that such pages even count.
To set your mind at rest, there are only 3 pages in the spam category at the moment. Most people would say that 3 is small compared to 2500. However, I agree with the sentiment that the number of pages is a rather crude measure of the content of the n-lab. I'm not sure how to find a better metric, particularly one that can be automatically computed. As a slight improvement, we could take a count of all the pages that weren't in the spam, people, or meta lists (hmm, looking at the list of used categories, perhaps it would be better to say "any page not in a specific category should be counted").
But we keep meaning to use categories more, don't we? Even people contains some pages that you'd want to count, such as M M Postnikov. And then there are the pitiful stubs with no category ….
And those with many categories. Should we consider the set of nlab pages or a multiset ?
Right now, there aren't any pages with more than one category. (At least, I don't think there are.) But if we start using categories more to categorise pages by subject matter, then there will be.
(^_^)
1 to 18 of 18