7

This is an advance link rot warning because an umbrella site (https://vitrinelinguistique.oqlf.gouv.qc.ca/) is being created to host 2+ different resources of the OQLF.

GDT (118 hits) / BDL (190 hits)

Bonjour,

Quand des contenus contiennent des liens vers le GDT ou la BDL, va-t-il y avoir une redirection vers la Vitrine ou va-t-il y avoir « link rot » et il va falloir modifier un à un tous ces contenus ?[...]


Les services de consultations linguistiques - 2022-07-11 12:34

Bonjour,

Il y aura des redirections qui seront faites à partir des URL des anciens sites (GDT et BDL) vers la Vitrine linguistique. Vos liens demeureront donc fonctionnels. Toutefois, ces redirections ne pourront pas être maintenues dans le temps. Elles dureront quelques années, puis les serveurs seront fermés et les redirections ne fonctionneront plus.

Pour un tour d’horizon des actualités concernant les outils, les services linguistiques et d’autres ressources de l’Office québécois de la langue française, abonnez-vous à ses infolettres en remplissant le formulaire d’inscription qui se trouve sur la page d’accueil du site Web de l’Office.

Meilleures salutations,
L'équipe de la Vitrine linguistique
Office québécois de la langue française


So how will the site be dealing with this?

3
  • If the links themselves change (and not only the base URL), I don't see any automated way of dealing with this. Let's build an army of little ants to do all the work!
    – Reyedy
    Jul 19, 2022 at 9:02
  • @Reyedy Updating the GDT urls seems easier to perform, whereas the BDL's include the section+subsection component before using the name of the article. I also just noticed in some cases they changed a bit the presentation, and adapted some examples to new realities or made them clearer, consider "Audrey est quelqu’un d’exceptionnel" vs. "Ma collègue est quelqu’un d’exceptionnel". Jul 20, 2022 at 7:23
  • Ouch, then I really think it's going to need a manual work. But to be fair it looks to be for the best! What we could manage to make automatic, though, is to scan all the existing links so that we can list them and follow the progress of the task! I know Glorfindel is very good with network-wide scripts, maybe some of his work could be adapted.
    – Reyedy
    Jul 20, 2022 at 13:23

1 Answer 1

5

First, thanks a lot for noticing this upfront and posting this question. This is usually noted after the links get broken and then it's much more difficult to fix them.

Second, this does not only affect French Language Stack Exchange, it's a network-wide problem:

This counts 13 hits twice, so we're talking about 552 posts here.

Now, what can we do? The first thing that can be done is archiving the linked pages into the Wayback Machine. That makes it easier to verify later what the correct new links are (if there's no "find and replace" type of fix). I wrote a script for that when a popular image hoster announced its shutdown; I haven't used it for a while but I've updated it now and the results can be found here. (Yes, the irony of using an external link is not lost on me - this is a consequence of the character limit of Stack Exchange posts ...) You can see the problems of using scripts for mass repairs at the bottom - rate limiting. But it looks like most of the content is still available, which is a good sign.

From your examples, I think a simple find and replace won't be possible. If http://bdl.oqlf.gouv.qc.ca/bdl/gabarit_bdl.asp?id=3381 would be hosted at e.g. https://vitrinelinguistique.oqlf.gouv.qc.ca/bdl/gabarit_bdl.asp?id=3381, that would be something staff could easily do with a script that doesn't bump questions; an example of this is all the edits by the Community user which change http:// links to https://. The ID of the article, 3381, is again visible in the new URL (though preceded with a 2), but just https://vitrinelinguistique.oqlf.gouv.qc.ca/23381 doesn't work; it seems you have to specify the full URL. (This is unlike Stack Exchange; https://french.meta.stackexchange.com/questions/896/ redirects to this very question, even though the full URL is different.)

As @Reyedy indicates, I have a script that repairs broken and outdated links across the network. It is capable of following redirections, and the response from OQLF indicate the redirections will be active for a few years. That should be enough time to fix all the links. There are two disadvantages of my script:

  1. it makes edits in my name, which bump posts to the front page (hence I'm always careful to rate limit it)
  2. I don't have enough reputation to edit posts (at least in this community), so the suggested edits have to be reviewed. The upside of this is that other users can make additional improvements to the post (grammar corrections, formatting). I guess I could try to amass enough reputation before the redirections are in place ...
2
  • Thank you very much, I commend your methodical, sensible and careful approach to the issue and I greatly appreciate you taking the time to write about it. If you can find the extra time at some point to post an answer to some Q on the main site, I can use a bounty to bump you up to 1k to enable editing if that helps. There is time and using you suggestions I'm sure we can come up with a plan. Cheers. Jul 21, 2022 at 0:29
  • @4469766572736974C3A9 I've updated the post with some results about the Wayback Machine - it seems we'll eventually be able to repair everything. As for your offer, that's very generous but I'm not sure that's allowed; one should vote for the content, not the user and awarding bounties to enable automated actions has been frowned upon by staff before. I'll have a look if I can write some decent answers, but haven't used French that much since high school :)
    – Glorfindel
    Jul 23, 2022 at 16:42

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .