PSA: LiveJournal lays off nearly half of its SF employees

[pne]

LiveJournal has laid off several of their employees in San Francisco (13 let go, 17 staying on seems to be the most reliable number I've heard), ostensibly to save costs by consolidating some of the functions in Moscow rather than having separate staff in Russia and the US.

This entry has a fairly good summary and links; if you haven't already, I encourage you to read at least the "onsite commentary" links.

While it seems that your journal is not in danger of going away anytime soon, a backup is always a good idea; a couple of the links in the "onsite commentary" have links to various backup methods, either in the entries or in their comments.

Comments

[ghewgill.livejournal.com]

*plug* ljdump (http://hewgill.com/ljdump/) :)

[pne.livejournal.com]

Which doesn't work if any of your userpic keywords contains non-ASCII text, as I found out when I tried it! (You get an "unhashable instance" error message, or something like that.)

I had ljArchive installed on my machine (from about a year ago, apparently), and that seemed to sync so I haven't put much effort into it.

At a guess, that the only thing that's needed is to replace r['pickw'] with asText(r['pickw']), where asText is (in pseudocode) "return list of [for item in list: IF item is an xmlrpclib.Binary object THEN item.data ELSE item]". (Presuming that the xmlrpclib.Binary objects use UTF-8 as their internal encoding.)

Though there might be other places where XML-binary-blob-to-string conversion will be needed.

(Also, the last time before that that I tried ljdump, it didn't pick up new comments on later starts if you had many -- something about incrementing an index by 100 or so until it's larger than the largest comment index LiveJournal gives you, then storing the last index away for later use and picking up there. Except that if you have, say, 234 comments, the index will be 300 at the point where it's larger than max-index, and trying to resume at 300 is wrong. I didn't check whether that was still the case but a quick eyeball of the code didn't look like the problem I had had back then.)

[ghewgill.livejournal.com]

Yeah, somebody else reported that (http://community.livejournal.com/ljdump/339.html?thread=10067#t10067) too. I'm going to have a go at fixing that tomorrow I think.

[pne.livejournal.com]

Ah, thanks!

Does the comment index thing ring a bell? Is it a known bug that you fixed in the past?

(I don't know what version of ljdump I encountered that on, nor exactly when, except that it was quite a while ago.)

[jixel.livejournal.com]

thanks/gulp

tho i'm sure they'd give us plenty hedzup if anything to happen. cmon now they're not evil like m$oft

[pne.livejournal.com]

You mean like the four days' notice they gave staff before they laid them off?

[pne.livejournal.com]

But yes, I don't think the site itself (and your entries) are going anywhere just yet.

But backups are always good.

[ghewgill.livejournal.com]

It's been a long time for me too, and I don't recall whether I had or fixed that comment index problem in the past. It sounds like the kind of problem that could easily happen with LJ's screwball API, though. :)

[jixel.livejournal.com]

well layoffs are layoffs (though i'd be anxious if i'd ponied up for a long term membership now). most likely the company itself would try and sell off its memberbase to some other co & get some $ for their assets, not just toss it all away.

[jixel.livejournal.com]

no idea if i want to mess with getting python to work on any of my puters. u done this? i joined the project's group tho so will checkitout.

[pne.livejournal.com]

I haven't tried ljdump as it doesn't handle userpic keywords with non-ASCII characters yet (and I haven't tried to fiddle with it, either).

I used ljArchive, which is a .NET GUI program, and it seemed to, more or less, Just Work. (Though I haven't looked at its output yet.)

[jixel.livejournal.com]

also weird download/go. i would think just a really smart rss reader could pull all our shite offline, mais non?

[pne.livejournal.com]

I think LiveJournal's RSS feeds only provide the last 25 entries.

So, if you have only 25 entries, then yes. Otherwise, no. (Even the smartest RSS reader can't read entries that the site doesn't provide in its feed.)

[jixel.livejournal.com]

ah. is that how it works? i could never understand why some sites have such a short history while others fling entire history back to adam.

maybe lj could offer a backup option ($) seems they need the money!

[pne.livejournal.com]

maybe lj could offer a backup option

They do (http://www.livejournal.com/support/faqbrowse.bml?faqid=8). It's rather inconvenient, only going a month at a time. (And the comment backup system isn't particularly pretty, either.)

And a "proper" one would have to be developed and tested first. (Not to mention supported.) Which, in turn, would cost money. So I'm not sure whether they'd even offer one as a paid option.

There's also jbackup.pl (http://code.livejournal.org/svn/livejournal/trunk/src/jbackup/jbackup.pl), which comes with the core LiveJournal code. (I haven't tried it out, though.)

So LiveJournal already provides *a* backup option. Just perhaps not what Joe Random User would like.

oh cool

[jixel.livejournal.com]

well actually that once a month thing is cool. i've only been here a couple yrs i spose so i could live with that for now. i have blog posts on other sites i am copy/pasting to others so at least here i can snarf a month and store moving fwd. but yeh commentses would be nice to keep too, not to mention saved memories of anyone's posts.

that perl thing looks scary too, but at least getting closer to my skillz

joe random user

[jixel.livejournal.com]

in other news, did you hear j.t. plumber's got a commentator gig now?

seriously

exporting comments

[pne.livejournal.com]

but yeh commentses would be nice to keep too

There's http://www.livejournal.com/developer/exporting.bml, but it's more for developers (as the URL implies) than for casual users.