How to restore a deleted/tampered question?

Sometimes users remove major parts of their question, which makes the answers/comments much less useful for future readers. The reasons why people do this could be discussed, but don't matter here.

Because people agreed to the Terms of Use, they published their post under the Creative Commons Attribution Share Alike 3.0 license. This gives anyone permission to post the exact content in a comment (with attribution of course, although that should be obvious from the context).

[Rik]

I wonder if contributors with editing privileges should attempt to restore edits. If there were no deleted files and the Google cache is complete, I don't really see a reason to wait until Rena (or another Answers dev of course) has time to restore edits.

[Adam Danz]

Wondering the same thing.....

[per isakson]

These are the current titles of two recent questions. Both have answers.

• [THIS QUESTION HAS BEEN REDACTED]
• CASE IS CLOSED **

They show how to "delete" questions. Even if I restore the question to a comment the title is still very visible.

[Rik]

Hence my question why we shouldn't edit the question ourselves. For the first one there didn't seem to be a capture of the cache yet. When I'm back to a pc I'll edit my comment to actually contain the original question. (requesting the desktop version of the Google results allows you to even follow the guide below on mobile)

Or do you mean something else?

[Matt J]

I wonder if contributors with editing privileges should attempt to restore edits. If there were no deleted files and the Google cache is complete, I don't really see a reason to wait until Rena (or another Answers dev of course) has time to restore edits.

One main reason I personally haven't been doing so is just that I didn't know the art of recovering the orginal post from cache. You've addressed that.

However, an additional problem is that determined OPs will just delete the question again. When you post the original version in a comment box, they can simply bury it by posting an unrelated thread title and 10 pages of SPAM in the question box, like this particular OP spectacularly did. What is needed is a way not only to recover the original question, but also to lock it from further editing. The editor community has put in a request for such a feature.

Answer 1

[Rik]

As an example, I'm going to use a thread that was recently edited: https://www.mathworks.com/matlabcentral/answers/665533-how-can-i-plot-this-type-of-figures

Remove everything after the thread ID and paste that in a Google search:

That gets us these results. You can get to the cache by clicking the arrowhead next to the URL.

First confirm we are looking at the correct thread. We searched for 665533, so that should be in the URL in the cache as well.

This checks out: we are looking at the correct thread. Unfortunately, this cache is from after the tampering.

Returning to Google, we see the UK mirror shows up as well. Sometimes you need to click the ‘repeat the search with the omitted results included’ button at the bottom of the page to see all mirrors. In this case clicing that button I can find another result from the UK mirror, but also the Swiss, Italian, German, and Korean mirrors. Sometimes these are updated more slowly, giving you multiple chances to find the latest version before the tampering. Sometimes the arrowhead is missing, which means there is no cache available for that version of the page.

In this case we can use the UK mirror that was the second result.

The thread ID matches, so this is indeed the correct thread. It is also from before the tampering.

Now let’s create a permanent archive, because the Google cache will be updated at some point to reflect the live (and thus tampered) page.

I prefer to use the Wayback Machine, which you can access on web.archive.org. You can prepend http://web.archive.org/save to the URL, but you can also navigate to web.archive.org/save to get the page below.

Paste the Google cache link in the box (optionally uncheck ‘Save error pages’) and click the ‘SAVE PAGE’ button.

Now you have a permanent copy of the cache, which you can use to copy the original question content from.

Don't forget to put that link in your post, as that is the only easy way to get to the saved cache.

[Rik]

This could probably be automated, but that would probably require a Google API key.

[Adam Danz]

Thanks, Rik!

I haven't used it (yet?) but there's a chrome extension for easy retrieval-from and archiving-to the WayBackMachine

https://chrome.google.com/webstore/detail/wayback-machine/fpnmgdkabkmnadcjpehmlllkndpkmiak

[Image Analyst]

For what it's worth, I visited the Internet Archive in 2017. We just looked at it from the outside at night after they were closed. I'm attaching photos.

[Adam Danz]

That must be the delivery truck that delivers fresh shipments of archival data to the facility. 🤪

[Stephen23]

Important Addendum:

After completing the steps given in the above answer, the MATLAB Answers thread itself must be added to the Internet Archive, otherwise it is easy for those links to be removed and all of that effort is in vain.

For example, an OP deleted all of their questions and comments and then all of the archive links that volunteers added to recover that data after the original questions were rudely removed:

https://www.mathworks.com/matlabcentral/profile/authors/13480323

If the thread itself had been archived immediately after adding the archive links, it would be trivial to recover them later.

PS: I just added this page to the internet archive.

[Image Analyst]

It seems that this is only good for adding a page to the archive if you can get an old copy already from an archive. But the times I've tried, usually a day or so after the page was altered, either the various archives don't have the page at all, or if they do, it's of the already-altered page.

And there are so many first time posters I can't spend the time to add all of them to the archive myself.

[Rik]

I would estimate my success rate at about 95%. I don't keep track of any real statistics, but I admit I tend to try it within a few hours (if it takes longer Stephen tends to have it done already). I did resort to the Bing and Yahoo caches once or twice to find an unaltered page.

I don't think we should archive each and every page to the WBM. I think it is on Mathworks to implement one of the various suggestions (block edits removing >80% of the content, create a restore point each time an answer is posted, etc).