A bulk post shredding tool for LJ

[crazyscot]

As a visual protest, I wrote a tool that bulk-shreds the contents of an LJ.

In case anybody wishes to join me, I have shared the tool. ljshred@github | just the script. You'll need Python 2.7. Command-line only. No warranty. Use at own risk. May contain nuts.

I may or may not delete my LJ now that I have run this, but for the time being you can see the effect over on http://crazyscot.livejournal.com/ .

Comments

[shadowkat]

Oh, you lovely person. I applaud this. Wish I'd been able to do that, instead I just figured out to delete mine without signing the agreement.

Right there with you.

[crazyscot]

It should still be possible to log in to LJ and undelete your journal within 60 days.

[shadowkat]

Oh, I know. Thanks for the suggestion all the same.

But, no, it's staying deleted. Not really techie or have the patience for it.

Besides, I was on way too many social media. This gives me a welcome excuse to get rid of one of them.

[filialucis]

How did you manage to delete your LJ without agreeing to the new ToS? I'd like to be able to do that myself. It's outrageous that I've got to agree to the things if the reason I want to delete my journal is because I DON'T agree with them!

[crazyscot]

Disable JavaScript in your browser.

[filialucis]

Thanks!

[shadowkat]

You turn off the "scripts". I installed Ghosterly, then disabled the tracking scripts on LJ, which in turn turned off the
the user agreement pop-up. It doesn't give you a lot of flexibility, the most I could do was delete my journal.

[omnipotent]

Wow, that's awesome. Nice!

[mdlbear]

Nice! Not quite what I'm looking for -- I want to replace posts with a link back to the archived version on Dreamwidth -- but I really love the concept.

[filialucis]

Now that would be a neat thing to do. Is it even possible?

[crazyscot]

Certainly looks doable to me. Through their APIs you can download posts from both sites, match them up, then replace LJ posts with links to DW.

[filialucis]

Uh... *looks blank* :D

Thanks, but I think I'll take a pass then! No clue how I'd put that into practice; probably quicker just to delete the LJ.

[hazard]

quite brilliant

[gehayi]

Brilliant. And beautiful.

I'm going to have to have an LJ account for another week or so (until I can import some comms to Dreamwidth) but after that, can I please gack your reasons-for-deleting letter? I'll credit you.

[crazyscot]

Of course!

[gehayi]

Thank you!

[sapphiren]

Awesome thing. However, will they be keeping original posts on their servers?

[crazyscot]

A very good question. I suspect not, knowing something of how databases are programmed, but I have no evidence to back that up.

[sapphiren]

Thank you!

[marahmarie]

When I saw "shredded" I figured "deleted", "blanked out", "gone" - then I saw it. Neatest thing ever. Looking at it, I thought: "base64", "something based on binary", and of cryptography in general, before it hit me: "Oh, so that's what they mean by "shredded"!"

My question is, can it be reverse engineered, or is all the text forever gone (and edit: I don't mean "can this be reversed" - I read the instructions. I mean, literally, "can it be reverse engineered"? There is a difference)?

[crazyscot]

The text is completely replaced with blocks, or randomly-chosen garbage, at your option. The shredded entry contains nothing of the original text.

Technically, a tiny amount of information is retained: in both cases I leave spaces as they are, so the entry looks "word-shaped" and you can quickly figure out the lengths of the words, but there are no further clues as to what the letters and punctuation are. I don't personally think this is enough to worry about, and the primary goal of my tool was the visual effect, not full information-theoretic security.

[marahmarie]

Thanks for clarifying. Anyone who has their LJ set to be indexed by search engines will still leave behind caches that defeat using this for security (at least until the caches are updated to reflect all the junk data, but from my experience Google will focus on the most frequently touched/accessed pages and many other pages approximately never) but assuming indexing was not allowed to begin with on a journal, this does look and sound fairly hard to reverse engineer for the average, non-determined person. And awesome to look at, besides. :)

[crazyscot]

I'm not sure how I found this out, but I once discovered that all my public LJ posts had been duplicated by some random bot and put up on a server that was largely Russian-language. Most peculiar.

[marahmarie]

Yeah, that does sound kind of weird, though not too surprising - if you had your LJ indexed by search engines and it was placing well for keywords the Russian site wanted to place well for, that might have been why - Google has a strange way of handling that, or at least did when I used it from 2005-2010 - it would rank sites that used underhanded SEO and crazy amounts of backlinks to one another better than it would rank LJs, while blissfully ignoring that the content was scraped from an LJ to begin with (a fact that should have sunk it in search rankings pretty much instantly). I really don't miss those days.

[mc776]

Neat!

Any plans for a Zalgo effect or just replacing every word with "chicken"?

[crazyscot]

It's just a bit of Python. Replacing every word with chicken is trivial. Zalgo is a bit more work.

[crazyscot]

I like both these ideas, so I've just implemented them. Thanks!

[mc776]

Awesome!

(I almost wish I hadn't deleted and purged my own LJ a couple years back...)

[paw_prints]

Oh wow. I WISH I'd found this post before I spent two days deleting twelve years worth of entries!!

Good on you and thanks for sharing!