11
$\begingroup$

During my reviewing of Suggested Edits, I have encountered several of them attempted by user @Glorfindel. Upon closer examination, all were performed by a robot written by him and all consisted of replacing inlined "http://" images (which are no longer displayed, being considered insecure) with "https://" ones. While this might be a good idea, my question is: shouldn't this be done on a wide SE-basis automatically by some SE software, as it was done two years ago for links internal to SE? Should non-SE software be allowed to do the job, at least as @Glorfindel describes it?

Until having some majority opinion on the issue, I shall reject these edit attempts. If the community decides that these edits are fine, then maybe they should be performed with some form of higher privileges, in larger numbers (not just 3 per day, as it happens now), and maybe not as "regular" edits.

$\endgroup$
7
  • 1
    $\begingroup$ Just to clarify, you're asking specifically about MathOverflow right? (If not, shouldn't this be raised on Meta Stack Exchange. In fact, I'm pretty sure some related topics have already been discussed there.) $\endgroup$
    – Martin Sleziak
    Commented Jan 2, 2019 at 20:11
  • 23
    $\begingroup$ I'm not sure I understand the issue: the "robot" is not some autonomous/anonymous agent, it's a script by a user who remains accountable for suggested edits; if these are valid edits, why reject them solely because they were carried out by a script rather than by direct human intervention? $\endgroup$
    – Carlo Beenakker
    Commented Jan 2, 2019 at 20:46
  • 14
    $\begingroup$ I agree with @CarloBeenakker: the method by which the edits were done is essentially irrelevant compared to the edits themselves. As long as the frequency is low enough to avoid flooding the front page, I don't see any problem. $\endgroup$
    – Denis Nardin
    Commented Jan 2, 2019 at 21:13
  • 3
    $\begingroup$ @CarloBeenakker: If the edits are deemed useful, they should be taken over by the SE staff and be performed by the user Community (possibly on a network-wide scale), as has been done in the past in similar circumstances. $\endgroup$
    – Alex M.
    Commented Jan 2, 2019 at 21:37
  • 4
    $\begingroup$ Probably it goes without saying, but if a post is bumped for some other reason (in this case dead link to an image), it is a good opportunity for MO users to improve also some other aspects of posts which might need improving (e.g., tags, grammar, typos, references, links to papers, etc.) At least, it is definitely better if more stuff is improved at the same time than editing and bumping the same post for some other reason a few months later. $\endgroup$
    – Martin Sleziak
    Commented Jan 3, 2019 at 11:04
  • 1
    $\begingroup$ ... it reminded me of a story told by a friend, where a manager forbade programmers working on an Y2000 porting project to use vi's macros, insisting that they must edit everything by hand... I think it smells of an Luddite approach, forbidding automated edits like this. $\endgroup$
    – Dima Pasechnik
    Commented Jan 8, 2019 at 18:56
  • 2
    $\begingroup$ ... and my recent experience with buying tickets on site of Scoot, a Singapore-based airline, which at some point started popping up captures and asking to complete them, telling me that it suspects I am a robot, as I type/click too fast... I'll go home now to check whether I'm still a human being ;-) $\endgroup$
    – Dima Pasechnik
    Commented Jan 8, 2019 at 19:01

1 Answer 1

Reset to default
34
$\begingroup$

Just to give some background on this project: it's indeed something I'm doing on my own, independent of Stack Exchange. There are more independent 'volunteer' projects around the Stack Exchange network, e.g. Smoke Detector (which I'm also a member of) which automatically flags spam across the network (including MathOverflow), and a dozen bots performing various moderation tasks on Stack Overflow. Some of these tasks could be done by Stack Exchange the company, but they simply lack the budget and/or (wo)manpower to do so. I'm not asking for permission from every single community, as that would amount to me having to post 173 questions on each single meta site. I've pointed out some means to contact me on the GitHub page, but this Meta post will do fine; thanks @MartinSleziak for notifying me about this post.

You're right that as part of the transition from HTTP and HTTPS, all HTTP links to sites inside the Stack Exchange network were automatically rewritten to HTTPS by a script made by Stack Exchange employees. However, that was a relatively simple script; it did not check whether the target still existed, or could instead be retrieved via the Wayback Machine.

One of the nice things about that script was that it didn't bump old posts; mine does (which is good, so other users can (re)view what it does), but there are other scripts that bump things (recent example). If this community finds 1-3 bumps per 36 hours too annoying, I can decrease the frequency. Other conditions can be discussed; if, for example, the bot should only attempt to fix posts with a positive score, I can try to program those conditions into the bot. Switching it off entirely for MathOverflow is also an option, but to me that would be equivalent to a government ban on restoration of old paintings. For what it's worth, at this moment, 97.5% of the suggested edits by the bot have been approved (network-wide).

For the record, there are currently 41 posts left on MathOverflow which the bot will try to fix.

$\endgroup$
13
  • 3
    $\begingroup$ If instead of doing three at a time once every 36 hours, it did one at a time spaced evenly, I would prefer that. Also, if there is some minor indication that it is an automated edit (GlorfindelbotA?), I could ignore the post instead of looking for a substantial change in content, especially on questions where I am looking for fresh content. Gerhard "Willing To Coexist With Bots" Paseman, 2019.01.02 $\endgroup$
    – Gerhard Paseman
    Commented Jan 2, 2019 at 21:32
  • 15
    $\begingroup$ I think a bump every 12 hours is harmless and is doing a desirable service. I agree it would be better if SE did this (and did not bump in the process) but I do not see them champing at the bit to do so. $\endgroup$
    – mme
    Commented Jan 2, 2019 at 21:54
  • 6
    $\begingroup$ @Glorfindel, given that SE is a community-centered project, why not make the code of your robot publicly available (for instance, in a GitHub repository), under some free software license? $\endgroup$
    – Alex M.
    Commented Jan 3, 2019 at 8:58
  • $\begingroup$ @Glorfindel: The fact that 5% of your edit attempts got rejected reinforces my thesis that this kind of massive, large-scale edit project should by performed by some other, more privileged channel, and not by regular edits (which require the users' approval). $\endgroup$
    – Alex M.
    Commented Jan 3, 2019 at 9:03
  • 13
    $\begingroup$ 95% was an estimate, it's actually 97.5%. Half of the rejections were because of "clearly conflicts with the author's intent", which says more about the reviewers not paying attention than the edits themselves. $\endgroup$
    – Glorfindel
    Commented Jan 3, 2019 at 9:40
  • 4
    $\begingroup$ I'll think of publishing the source code. There are some problems with this: 1) when multiple users are running the program simultaneously, you'll lose control over the rate-limit (not to mention the errors which occur when you're trying to suggest an edit to a post which already has a pending edit). $\endgroup$
    – Glorfindel
    Commented Jan 3, 2019 at 9:44
  • 2
    $\begingroup$ 2) you'll need a registered application on StackApps to run it; 3) you'll need that app to generate an access token to identify your account. So it's a quite convoluted process. $\endgroup$
    – Glorfindel
    Commented Jan 3, 2019 at 9:46
  • 7
    $\begingroup$ With the edit, it becomes clear that at the current rate this will be done in about a month. So is there really any harm? $\endgroup$
    – mme
    Commented Jan 3, 2019 at 16:23
  • 3
    $\begingroup$ There is something that puzzles me in this connection: I have many old answers that contain an inline image referred to with http rather than https; they seem to display just fine; for example: mathoverflow.net/a/143866/11260 --- any idea why they do not suffer from the http ban? $\endgroup$
    – Carlo Beenakker
    Commented Jan 3, 2019 at 22:07
  • 2
    $\begingroup$ I don't know; it might be because you used capitalized HTML tags; it might be because it's old and 'untouched' for a long period. Here is a post with similar markup, a link which is still alive and yet an image which is not working. I've added some extra precautions to the bot to make sure it won't try to fix images like this. Pending the outcome of this discussion, I've also disabled it temporarily for MathOverflow, but I'm humbled by the support I'm getting. Thanks! $\endgroup$
    – Glorfindel
    Commented Jan 4, 2019 at 8:23
  • 1
    $\begingroup$ thanks for the explanation; I will slowly and manually (no robots here :) make the http --> https replacement on my old posts containing inline images; hoping to avoid annoying anyone with bumps of old posts. $\endgroup$
    – Carlo Beenakker
    Commented Jan 4, 2019 at 13:19
  • $\begingroup$ It's two weeks later and many people have seen this thread (partly due to being a Hot Meta Question). Given the reactions and upvotes, I'm going to re-enable the bot for MathOverflow tomorrow. $\endgroup$
    – Glorfindel
    Commented Jan 19, 2019 at 13:10
  • $\begingroup$ I wish it was only three bumps per 36 hours. I personally find the editing and bumping from yourself, which I have no doubt is motivated with good intentions, to be extremely annoying. I am sorry but that is how I feel. $\endgroup$
    – JP McCarthy
    Commented Dec 2, 2022 at 17:32

You must log in to answer this question.

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