Poll: should moderators edit posts to update hyperlinks?

Problem

One drawback of improving user documentation is that changes to the paths of statically-generated HTML can break links to the documentation pages.

Current behavior

Broken links to Purism resources currently results either in a 404 error (blog posts), or the user is forwarded to the root directory (user documentation).

Question

Should moderators be allowed to edit user posts to update hyperlink paths?

Conditions

  1. Posts will be edited only if one or more links in the original post is outdated or broken
  2. Post will be appended with the text [moderator edit: updated link]
  • Yes; for any hyperlink
  • Yes; only for Purism targets (blog, documentation, etc.)
  • Yes; after receiving user consent
  • No
  • I have no strong opinion on this matter; just show me the results
0 voters
3 Likes

I take it redirects (for instance using a .htaccess) canā€™t be used (301 or 307 HTTP response status codes - HTTP | MDN) on the documentation end?

2 Likes

As long as condition 2 is respected.

2 Likes

I think thereā€™s a bit more to it as well. If the URL of a page is strictly renamed (moved) and itā€™s a Purism page, I see nothing wrong with updating the URL. But what if the documentation is restructured so that what the original link linked to is not exactly the same general content as what it now links to? That makes it more of a judgement call as to whether itā€™s appropriate to change the URL.

Of course the entire web suffers from that problem. Link rot, and change to the target page, are problems throughout the entire web. This link degradation is unhelpful everywhere.

Bottom line though is that within the narrow scope of Purism forum links to Purism documentation pages, it is helpful that links continue to work.

4 Likes

The only downside that I see there is that over long periods of time your web site will become an unmaintainable mess, as you try to preserve the validity of every URL that ever existed by redirecting to a current URL.

This may be even harder if content is being generated automatically from some source and/or is being served out of a content management system. (The latter is a two edged sword because obviously it gives the potential for automating some of the behaviour but only if the CMS can actually do what you need it to do.)

It may even become impossible if a long obsolete URL ends up being valid again under a different purpose.

4 Likes

If a user included a header anchor, that could be used to locate specific documentation - otherwise, it is very generic and could be subject to (mis)interpretation. Context clues are important in this case.

2 Likes

Yes, starting with @david.hamnerā€™s posts containing Matomo Tracking Module URL parameters.

Perhaps that should be automatic behaviour across the whole forum.

2 Likes

austin-powers-dr-evil-3118736152

2 Likes

Especially when they paste the unnecessary stuff AFTER a question mark. which probably tracks you for adverts.

1 Like

I didnā€™t see any mention of how the Moderators would feel about the extra tasks of checking links, hourly, daily, weekly, monthly? Hmmm? Or do visitors and members need to report them?

How many would bother to point out the bad link?

Does Discourse have a option to view errors in itā€™s own system? Is it possible to make Discourse test links and then re-linking or editing the post with [this link is dead] notification beside it?

From the number of errors I get, Iā€™d suggest that all the help information and How-toā€™s be listed on one page, as a Index so to speak.

My side beef with errors is I have found it more efficient and accurate if I go to a Search Engine, or Google and search for/google exact phrase.
So I tested the following:

Information needs to be corrected on Posts, Products, and Shop pages

I will get:

This site canā€™t be reached
The webpage at Information needs to be corrected on Posts, Products, and Shop pages
might be temporarily down

Note that # near the right side. Thatā€™s why search engines and Google get a error.

What I do, when I get that 404, is to go to the end of the URL and add /1 and since there is a post, there must be a 1st post and I get to the Topic page containing the related pages.

The above is a example. But I stepped X stepped through it to make sure.

I do that with (/1) almost every search starting with ā€œlibrem 5ā€ and search keywords. Google is getting better, itā€™s the search engines that choke on it.

Too, some external links within a post that sometimes donā€™t count. (the numbers that indicate how many clicks the link gets).

To the question, No, why burden the Mods with extra work?

2 Likes

I took the question as giving moderators the right but not an obligation.

In other words, if a moderator sees a broken link and the moderator would be motivated to rectify the link, is the original author of the post, generically speaking, going to get his or her nose out of joint if the moderator were to go ahead and fix the link? It doesnā€™t mean that the moderator is going to go ahead and fix the link in any given case.

Also

I donā€™t know whether Discourse specifically can do this but some systems do have automatic periodic broken link detection. So if the consensus is that fixing broken links is OK and Discourse has this functionality then at least that part of the job does not impose extra load on moderators.

2 Likes

I am often tempted to replace a broken link in somebody elseā€™s post with a link to Web Archive if I find an archived copy of the page. It is a habit from Wikipedia and Stack Exchange.

For checking a website for broken links, LinkChecker works great. However, I havenā€™t tried it with Discourse. I confirm that manually correcting all broken links is an undertaking.

3 Likes

To clarify my previous post ā€¦ in saying that some systems do have broken link detection, I am talking about internal to the content management system / document management system / whatever exactly produces the web pages ā€¦ which is typically much faster, and imposes no network load, doesnā€™t make the web server logs a bit bogus, etc. etc. ā€¦ but not quite as definitive.

You can always check a web site from outside (subject to the usual caveats and, these days, even tar pits :wink: not that that applies to this forum).

2 Likes

Ah yes, the fine print :grinning:

I was just checking Discourse and itā€™s not a thing they have although it is talked about a lot. (broken link detection/repair/mark.)

Out of curiosity, can you tell us what or how you run Discourse? On a VPS of course, and as 2015 does not work with any of the GUIs like ā€œcpanel, plesk, webmin, etcā€ what do you use for the other websites like docs, store etcetera.

~s

2 Likes

Was that question directed at me specifically or Purism or any rando who wants to answer?

I personally donā€™t run Discourse ā€¦ which is why I donā€™t know whether it has broken link detection. Iā€™m just a humble moderator, and I donā€™t have admin rights to Discourse here, so I donā€™t control that kind of thing anyway.

1 Like

Yes.

I was curious if Puri is using Discourse server, or on itā€™s own VPS - a requirement as I read it for self-hosting.
If we knew if itā€™s self-hosting or using Discourse server, one can read further in one direction or the other about what is available and what scripts may help manage broken links, if it is possible and I read that it is possible.

There are also some sites that monitor pages testing links and if one is broken someone, likely the administrator, is notified. But $'s matter at Puri which is a good thing, maybe.

As Paul Harvey would sayā€¦ā€œAnd now you we know the rest of the storyā€. :slightly_smiling_face:

You forgot ā€˜modestā€™ too. :rofl:
~s

3 Likes

Purism uses Hetzner for self-hosting Discourse.

1 Like

(OPINION)
Yeah, about Hetzner. I didnā€™t go through their frontline barrier that gives 1 of 2 options required to just read their site (1) give up all rights to privacy, or (2) give up all rights to privacy.
Again, itā€™s my opinion only.

So, I didnā€™t get far, and in my opinion, would not recommend Hetzner as a Host.

I have already read the how to install Discourse article before. I wasnā€™t looking to install it. A little more info about whether the Discourse used here is the free version, or the u$eful ver$ion. Knowing that, one could start looking for a plugin or whatever, that would help trap AI scraping.

~s

2 Likes