Martin Sleziak's room

1:51 AM

Thanks for the SEDE queries! I should know how SEDE works by now, and clearly it was an oversight to have http:// in my search string...

Martin Sleziak

7:22 AM

@CalvinKhor You might have noticed that recently there were some edits of such links by Glorfindel.

Must Basis of an Euclidean Space Be Ordered math.stackexchange.com/posts/5904/revisions

Irrationality of powers of $\pi$ math.stackexchange.com/posts/2284/revisions

Divisor -- line bundle correspondence in algebraic geometry math.stackexchange.com/posts/1926/revisions

And if you look at his recent activity, you can see similar edits also on several other sites.

I assume that the edits have been made by his script - but we might wait for him to confirm.

Looking at the post linked by @hardmath, I suppose that also the links containing planetmath.org/sites are broken. But in that case, I don't see any easy way to fix them. (Maybe some of them might be available in Wayback machine?) — Martin Sleziak 30 secs ago

Martin Sleziak

8:35 AM

BTW quite a long time ago I made a not to myself about broken PlaneMath links: chat.stackexchange.com/transcript/19138/2020/9/23 Clearly, I havent looked into this since the.

Now I noticed this on MO:

@Glorfindel, would it be appropriate to make a MMO post, to go with Is it time to replace links to the UCDavis arXiv frontend?, to invite other users to join your crusade to fix PlanetMath links? (And, thanks!) — LSpice 8 hours ago

@LSpice Martin Sleziak already started one on Meta.SE, since this affects the entire network. For my script, it's a simple find & replace, the UCDavis case is way more complex. I don't think we want to bump 40 questions at once, so I scheduled the script to run once every three days. — Glorfindel 3 mins ago

@Glorfindel From the comments I copied above I see that you have already added to your script also fixing PlanetMath links.

When you have time, maybe you would be willing to post an answer saying that you're script is fixing such link to the post on Mathematics Meta: Old Links to PlanetMath have Broken

And thanks for doing this! (I expected that we'll be fixing them manually.)

Glorfindel

@MartinSleziak done, thanks

Martin Sleziak

8:50 AM

Thanks, I appreciate that!

I suppose it would be possible to do cca 180 posts manually - but since you are able to automate it in some way, that's definitely better.

I do not want to bother you too much if your busy - and also I assume that those links might be much less straightforward to replace.

But if you're willing to have a look also at the links to arXiv frontend, I can summarize somewhere what I can say about the ones which need fixing.

We were told by Catija that once the links until April 2007 are fixed manually, the rest can be fixed by Stack Exchange: Is it time to replace links to the UCDavis arXiv frontend?

4

A: Is it time to replace links to the UCDavis arXiv frontend?

I've reviewed your request and - while I'd generally want to say "yes, we can do this" - in this case, I'm sorry to say that I don't think it will be possible - at least, not without some effort from you all first. I really do think it's important that the links on this site are functional, so I ...

However, this is much less straightforward than PlanetMath.

Only some of the links (the old ones) should be converted.

And it seems more difficult to check whether they were converted correctly:

in MO editors' lounge, 14 hours ago, by Martin Sleziak

The links https://arxiv.org/abs/q-alg/9712048 works. The link obtained by the naive replacement does not work: https://arxiv.org/abs/math/9712188 - but it does not return 404. (It returns a page saying: "Article math/9712188 in the Mathematics archive not found".)

In any case, several MO users already started to edit them manually: chat.stackexchange.com/transcript/10243/2022/1/2

Martin Sleziak

10:47 AM

3

A: Old Links to PlanetMath have Broken

Because the ideal tool for a job like this, the mass-replacement tool, which can repair posts without bumping them, doesn't work in this case, I have fired up a script (a derivation of the Broken Image Repairer) to repair these links on half a dozen websites across the network, including Mathemat...

> I'm open for objections or suggestions. Following Martin Sleziak's comment, I can have the script try to repair planetmath.org/sites links as well, e.g. planetmath.org/sites/default/files/texpdf/40203.pdf in this question has a Wayback Machine copy.

Excellent, thank you! Do you expect to be able to fix every such broken link, and if not can you put out a list of exceptional cases? I'd agree with fixing the /sites links too. — Calvin Khor 2 hours ago

You're welcome! I expect the /encyclopedia fix to be highly effective, since that is just reindexed content; for /sites I'm not sure, but I can do a dry run first and check which URLs can only be fixed manually (and post them here of course). — Glorfindel 2 hours ago

Calvin Khor

11:07 AM

@MartinSleziak I found that earlier by searching in this room for dead-link which you put in this chat message. I thought it a natural tag to have on meta, and put it on the math.meta question hopefully this makes it harder to forget about

Martin Sleziak

@CalvinKhor Yes, I have noticed the tag dead-link. In fact, I have mentioned it in the tagging chatroom: chat.stackexchange.com/transcript/3740/2022/1/14 (And I have also mentioned similar tags on some other metas.)

I guess that it is quite likely that somebody might make another posts with this tag in the next 12 months, so that tag will not be removed automatically. (See: What are the differences between tag pruning on main sites and their metas?)

In fact, I have in mind some question which I might possibly put on meta and where this tag would fit. (I have to think about that question a bit more before doing that. And also I'll post it when I have a bit less work IRL.)

Calvin Khor

11:27 AM

There are plenty of posts that one could go back and tag with it e.g. one on EOM links , imgur links and the earlier post linked to by hardmath. There are many different names or related terms (broken link, dead link, link rot, 404...) so a tag should be useful. Not to say that we should force the tag to stay by editing such a post

Martin Sleziak

Yes, I agree that there are several older posts where the tag dead-links would be suitable.

Of course, bumping all of them at the same time wouldn't be optimal thing to do.

In any case, the period is 12 months - at the moment we do not have to worry about the automated removal.

I will add link to Glorfindel's queries about such tags - tags with only one occurrence (which might be removed after 6 months, on meta 12 months): main, meta.

Martin Sleziak

1:14 PM

Does the script have a rate-limiting feature? — This_is_NOT_a_forum 24 mins ago

@This_is_NOT_a_forum yes, configurable. See here, except that these broken link scripts (I have a couple of them running) run once every three days. — Glorfindel ♦ 19 mins ago

> The program runs approximately once every 36 hours. Edits are limited to one suggested edit at the time, and up to three autonomous edits to e.g. Community Wiki posts; this both limits the effects on the front page, and does not increase the burden on reviewers too much. (I'll run the program more often on Stack Overflow because of the vast number of posts to fix there; flooding the front page is rather hard there.) If the previous suggested edit is still pending review, ...

> ... the program will skip that site. I'm an avid reviewer myself and wouldn't like to review hundreds of the same type of edits. Also, whenever I'm able to, I'm trying to manually review the edits to correct typos and improve formatting. Reviewing by other people helps finding bugs like this one where the program attempted to replace images in code. Hats off to the reviewers there!

Martin Sleziak

3:05 PM

SEDE now returns 24 posts with planetmath.org/sites.

The search for url:"*planetmath.org/sites*" now returns 24 posts.

A SEDE query where they are ordered by last activity (i.e., the questions that were bumped most recently are shown first): data.stackexchange.com/math/query/1474696/…

There are 15 comments with planetmath.org/sites: data.stackexchange.com/math/query/1451220/…

The posts where such links was edited away: data.stackexchange.com/math/query/1070272/…

Users who often removed such links: data.stackexchange.com/math/query/1468213/…

And number of removals per month: data.stackexchange.com/math/query/1468214/…

Users with most posts containing the planetmath.org/sites and Users with most posts where planetmath.org/sites was removed

For comparison, on MO there are only 2 such posts.

The above queries return only one instance where such string was removed from the post: math.stackexchange.com/posts/1942113/revisions

Martin Sleziak

3:25 PM

However, in such cases fixing the post doed not necessarily mean that the string planetmath.org/sites is removed from the posts - instead a longer URL to Wayback Machine is added.

At the moment, the following queries return no results. But I suppose that if/when some such edits are done, they should be included there. (Unless I have mixed up something.)

Posts: data.stackexchange.com/math/query/1177886/…

Editors: data.stackexchange.com/math/query/1468958/…

Per month: data.stackexchange.com/math/query/1468963/…

Glorfindel

5:21 PM

@MartinSleziak yeah, I read that post as well, I'll see what I can do. At the very least, I'm planning to add a feature to my script to do a dry run, collecting information about how many links can be replaced by straightforward substitution (and leave those to the CMs), how many links with a Wayback Machine snapshot, and how many need to be done manually.

Martin Sleziak

I should have expected that you plan to look into that.

@Glorfindel I should have expected that you have seen that.

Glorfindel

:)

Martin Sleziak

In this case I expect it to be less straightforward. Maybe there's no harm if it's done manually. After all, in such cases the editors at least have a look what can be improved also in other answers.

Glorfindel

That's true. There's nothing wrong with fixing things manually, it just takes more patience.

Martin Sleziak

I remember that very very long time ago I edited a few PlanetMath articles - and planned to edit more.

Now I am not sure whether I would be able to find the ones I have edited. (I vaguely remember some of them.)

Maybe it is one sign of getting old - that website I used to contribute to become obsolete.

Glorfindel

5:32 PM

@MartinSleziak would be doable with SEDE, I guess?

Martin Sleziak

I do not mean posts on Mathematics which link to PlanetMath. (For that, some of the queries above might help.)

Glorfindel

oh, PlanetMath itself ... sorry

Martin Sleziak

Yes, that's what I meant.

Glorfindel

I'm working on a 13" laptop right now and have chat windows reduced to 80% zoom - not reading so well right now :P

Martin Sleziak

Different devices certainly change how well site can be used. (I think I remember many people complaining about using the site or chat on mobile devices. I haven't tried that myself.)

BTW the fact that LSpice noticed your edit on MO and offered to help shows that also MathOverflow has some users who are keen to contribute to repairing dead links and improving old posts.

Anyway, time for me to get something to eat. Have a nice day - and thanks for dropping by.

Glorfindel

5:58 PM

@MartinSleziak yep, that's nice.

@MartinSleziak sure, enjoy your weekend!

Transcript for

$Mathematics$ Martin Sleziak's room