Thinking out loud

[susandennis]

I've wanted a client backup of my Live Journal (not the xml export - one that I can really use) for a long time. What I've always had in my little pea brain was:

• one web page for each entry


• including all comments


• saved in a folder on my computer


• automagially

I can actually do this manually, as it turns out. so I think it's doable.

I can get the URL that includes the post and the comments:

http://www.livejournal.com/users/susandennis/xxxxx.html?nc=yy

xxxxx= the number of the entry
yy=the number of comments (which, it appears, I can set to something very high like 300 and get all)

Then I save that URL as one, single web archive (.mht) file and everything is there - the look, the words, the pix, the comments, the live links - everything. To capture all the comments or most of them, this save thing would happen something like 48 or 72 hours after posting...

So I'm thinking something like every night, this batch thingie would go up and look and see what entries numbers have been captured yet and are now at least 72 hours old and then create the appropriate URL and save as a .mht file into a specific directory. Done.

How hard can that be? Hmmmm

Comments

[melodysk.livejournal.com]

Well if your name is Melody Shanahan-Kluth , we are talking rocket science LOL

Susan Dennis however ......

Thanks

[pheon.livejournal.com]

I have been after the same thing. My main dislike of the XML thingy is that it doesn't include the comments. I wasn't aware of the 'nc=' option.

So I may work on the same process. But from a Linux platform instead of Windows.

Re: Thanks

[susandennis]

I think the nc option may not be necessary after all.

It looks to me as if once you capture the entry number, you've got entry and comments at /xxxx.html

At least that's what I get here using S2 and IE.

I'm going to think about it some more and maybe post the thot to lj_nifty tonight and see if anyone knows anything. And, by the way, there's a link in lj_nifty right now to (finally) the new Semagic beta.

Re: Thanks

[kyrielle]

Alas, the nc doesn't work like that - and for popular entries you aren't guaranteed of getting all the comments by this methodology. The links that produce the 'nc' aren't there to change the display (the page ignores that option). They're there to change the URL, so that people who want to be notified when there are more comments will...because the link will turn to "unfollowed" again if more comments are added. (But, it won't if an equal number are added and deleted.)

All comments show on the comments page until you exceed some threshold. Then, some comments are only represented by their subject links and you have to thread into them; and you can end up with multiple pages of comments. A look at any of the really talky posts in lj_maintenance will show what this output can look like.

It'd be neat to see such a program, it's just going to be a bit more complex than a straight capture, I think. Of course, if your entries aren't likely to draw heavy comments - I know mine aren't! - then a straight capture would work fine.

[pookfreak.livejournal.com]

If you figure it out and are willing to share the secrete, please do let me know as I'd love to be able to archive in a format I can use.

[estis.livejournal.com]

Since the xxxxxx.html part of the URL's do not seem to be in sequence I don't see how you can do what you want. It would help if you knew how they are numbered and work from there perhaps. Should add that script writing isn't my strong point.

[susandennis]

Yeahbutt see I'm thinking you go into SusanDennis and check date. Scarf up the xxxx's for each date and then you have them...

That's what I'm thinking.

I'll see your script writing skills and raise-er- lower you one.