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.
Looking at some recent statistics, on an average day the nLab gets the following number of requests per page type:
It seems easy enough to do some simple optimizations, like
Regarding the last point I want to suggest changing the behaviour of these pages so that they don’t take 10 seconds or more to load, while still giving us useful information. Therefore I want to ask everyone: do you use the recently revised page? If so, what do you use it for? As a reminder, it currently gives a list of every page ever edited in the nLab, sorted by the date of the latest revision. It would be much less resource-intensive, for example, to simply list all the edits made on the nLab in the last month (say) and split this list into several pages, each showing 100 edits (say). Would this still be useful to the people who use it?
And I would ask similar questions about the authors page.
Yes, I believe we had this kind of suggestion mentioned by several people before. People typically use “recently revised” to check current activity. It should be sufficient and would be useful if that page only showed changes going back just a little bit into the past.
That would be great if you could implement that.
I think Zoran had a reason to retain the full version, which I can’t find now. Might it be possible to have both pages? How recent should recently revised include? A month?
I think Zoran had a reason to retain the full version, which I can’t find now. Might it be possible to have both pages?
I’ll wait for Zoran then before implementing this. If necessary I guess it could be possible to keep the old page as well.
How recent should recently revised include? A month?
I was thinking one month should be sufficient, but I’m open to suggestions (and this is always trivial to change later).
maybe one could have two pages that do what their titles suggest:
“recently revised” showing indeed only recent revisions
“all revisions” showing the entire edit history
I occasionally check the Authors page to update Contributors or to check on spam users. But it takes so long that I use it rarely. It would be great to speed it up, but ultimately I would want to be able to get both a complete list of attributed authors and a complete list of pages edited by each attributed author.
The internal search function is much better than Google search for some purposes; in particular, it actually does a regex search of the source. The last time we discussed this, we put a Google search field on the main page for casual users while maintaining the internal search for hardcore users. Possibly we could make the Google search more accessible (such as on every page) to encourage its use, but I would like the internal search to still be available.
We have disabled both Export and Recently Revised in the past for reasons of speed. Then we reenabled them when some problems were solved. Andrew would know more about this.
Yes, sometimes I need full revision history (e.g. in order to catch the list of stubs which I started some time ago in a spree of activity on a temporarily abandoned topic), but that option could be on a less publicised link if this helps. Classification of Urs, 5, would also be helpful.
Export was already disabled before, it’s OK with me if you do it again.
But, as far as search, I find Lab search much more useful than the google search, especially because it separates the occurences in page title and in the text and because it looks through the source and not through the output. In addition the Toby’s arguments about it.
I also usually prefer the nlab search to the google search, for the reasons already given.
I agree with Mike.
Right, the search will stay then… I’ll try to implement Urs’s suggestion (#5) when I get a chance.
What I really wanted to do was split the server resources so that different types of request went into different queues. In particular, I wanted to put show in one queue, edit and similar in another, and then the rest in a third. That way, they wouldn’t block each other at the server level.
I think that to do that one would need to have a family of alternative URLs, such as show.ncatlab.org, which all pointed to the same location (ie the current nlab server). Then a request to ncatlab.org would get diverted to ABC.ncatlab.org by Apache’s rewrite module. Each ABC.ncatlab.org would then be running a separate instance of passenger but the passengers would all point to the same instiki installation. This, I think, would relieve the main bottleneck which is the number of concurrent requests and the fact that a few big ones can block the others.
In all of this discussion, we should also recall that All Pages was disabled and (unlike Recently Revised) remains disabled. (But category listings with All Pages work, at least modulo the bug reported in a different thread).
The recently revised page should be much faster now, due to a caching trick I implemented. That said, it only updates every half hour now.
I replaced the “Recently Revised” link at the top of every page with a link to a new page called “Latest Revisions”, which shows the last 1000 revisions made (spread across 10 pages). The recently revised page is still there at the same link. Let me know if anyone encounters any issues with the latest revisions page (though I’m heading to a conference now so will probably not respond promptly).
It's fast! Even Recently Revised was tolerably quick.
Looking at Latest Revisions at 16.36 BST (GMT + 1) 1 Sept., the most recent entry is
Sandbox ( Sandbox(rev # 514) | history ) by Adeel Khan at August 30, 2014 23:52:10
I’m not sure what time zone that’s from, but it’s obviously not the latest.
Recently revised has disappeared from the top bar of Latest revisions.
Tim's link should be Recently Revised.
This also stops counting at August 30.
That’s embarassing… but the latest revisions should be fixed now.
It was fixed, but seems to have relapsed now to August 30 as latest.
That’s very strange, as it’s still up to date for me.
It was fixed, but seems to have relapsed now to August 30 as latest.
That’s very strange, as it’s still up to date for me.
On my machine, Latest Revisions is September 5, while Recently Revised is August 30.
(and why can’t I get a blank line between the two above quotes)
Right, the recently revised page is still broken. I’m on the way back home at the moment so that should be fixed soon.
Sorry, yes it was Recently Revised. And thanks for doing this.
Recently revised is now working again.
The category-specific Recently Revised pages (such as http://ncatlab.org/nlab/recently_revised/Galois+theory) are now broken. (They show all pages, not just those in the given category.)
Why all pages for given categories often point to spurious wanted items ? For example http://www.ncatlab.org/nlab/list/combinatorics gives that http://www.ncatlab.org/nlab/new/differential+cohomology+in+a+cohesive+topos (nonexistent entry) is wanted at Urs Schreiber, smooth infinity-groupoid, nonabelian cohomology, 2-group, topological T-duality, torsor, AKSZ sigma-model, abelian sheaf cohomology, cohomology, orbit category, twisting function, induced representation, differential cohomology, bundle gerbe, local system, motivation for sheaves, cohomology and higher stacks, Eckmann-Hilton duality, higher category theory and physics, string field theory, principal bundle, vector bundle, orientifold, string structure, Eilenberg-Mac Lane spectrum, tmf, gerbe, Hochschild-Serre spectral sequence, fiber sequence, principal 2-bundle, principal infinity-bundle, group cohomology, twisted cohomology, generalized (Eilenberg-Steenrod) cohomology, generalized cohomology, Sweedler coring, smooth space, algebraic K-theory, bialgebra cocycle, chain homology and cohomology, K-theory, Karoubi K-theory, supergravity C-field, cup product, concordance, twisted bundle, bundle gerbe module, twisted K-theory, topological K-theory, representation ring, Bredon cohomology, fiber bundle,
….etc etc
and, by my inspection, it does not seem to be true.
Zoran, this bug dates from Jacque's original introduction of categories. It occurs whenever a page in the category links to a page not in the category, whether or not that page exists.
whenever a page in the category links to a page not in the category
But those pages do NOT link to the quoted page, in category or not, nor that page exists. I suspect that they linked to it in the history, but the data are somehow retained.
Ah, I see. Yes, that is different.
Both All Pages and Recently Revised are broken with respect to categories. It would be very nice if at least one of these could be made to work.
Examples:
(And some day I'd like to figure out where we're getting our Markdown updates from too, so we could fix that; but that's a separate issue.)
OK, I figured out what's up with All Pages there. If you write category: empty, then it shows up; if you write category:empty (no space), then it links the list but does not itself appear. I guess that that's a feature, albeit an obscure one!
But Recently Revised is broken for categories. This should be allowed to work; indeed, it's much less taxing if it works properly!
That is certainly an issue which should be corrected. I wasn’t really aware of the categories feature when I started hacking around with the page. It’s on my to-do list.
1 to 33 of 33