Alternative blogging hosts?

- A big pool of other interesting users, communities, and syndicated RSS feeds, with convenient aggregation of their postings in "friend lists"

This is the hardest thing to replace: friendfeed is an approximation, but is not perfect

- Threaded comments, with automatic notification by email when I get any replies to my posts or my comments

notification is standard with blogger/wordpress. blogger definitely doesn't have threaded comments, and I'm not sure about wordpress

- Easy posting, automatic RSS feed generation, cuts so that I can write long posts and have only shorter summaries show up on the front page of my journal, and the ability to apply arbitrary CSS to that front page

if you roll your own blog using wordpress, all of this is easy. If you use blogger/wordpress.com, it's still quite easy to do, except that I don't know how arbitrary the CSS can be (I think you can pretty much edit all of the header template in blogger/wp)

- Multiple user icons as a less-annoying substitute for emoticons

no luck here:

- Tags and tag clouds

wordpress has both categories and tags: blogger only has tags. In either, you can generate tag clouds, either directly, or via plugin

- I haven't used this much, but the ability to control who sees and who can comment on my posts is also a plus

Both blogger/wp have moderation options, but it's all or nothing: you can moderate everything or nothing. Neither (as far as I know) has the ability to do fine-grained control of post viewing: vox does this, but then vox is not well suited for most of the other things you want

Wordpress, as an application that I run on my webserver, doesn't have threaded comments; I would assume Wordpress as a service uses the same code and also doesn't.

Having grown up on Usenet, I would have said that threaded comments were critical, but Making Light has some good arguments in favor of not-threading, and it does seem to work out okay there, so maybe it's not as big a deal as I think. Certainly, large conversations in LJ comment threads are pretty rare, and reference-by-name-and-comment-number (since, IIRC, wordpress and blogger both number the comments) works well enough.

For me, the ability to control access to posts (and transparently give my friends access) is pretty critical, though, and so far LJ is really the only thing that's even close to working for that.

It's not so important on a blog with as few comments as mine typically gets, but (as heavily commented blogs go) I'm not sure Making Light is typical in its civility and careful referencing of previous comments.

I'd compare, instead, say Fark and Slashdot. Both have huge and mostly inane discussions, but (when I look at them at all) I find Slashdot's subject lines and threading quite helpful in picking out a smaller number of comments that might be of interest.

Also, a note -- notification is standard with blogger/wordpress for the post author. I don't think that comment authors get notified of replies -- actually, they can't, because there isn't the notion of replies to individual comments in the system. That's a pretty big limitation, as those notifications are somewhat critical for keeping long-term conversations going.

yep, lj still has some compelling features. i suspect these will collapse with some kind of notification solution, but i'm not sure graph data will open up as completely as it has here.

here's a feed of posts by lj users who have no e-friends.

http://tucc.us/x/ljwc_goobers.xml

one thing to notice: creepy marketing sludge.

Yeah, I should have listed having more of a focus on connections between journals and less of a focus on simply provide a platform for individuals to broadcast their thoughts to the world as one of the things I like about LJ.

not that it matters (since DE wants something much closer to LJ), but Blogger has a comment notification system for commenters (I use it extensively). WP on the other hand has comment RSS feeds, which is the next best thing (I use that a lot as well). In both cases, commenters get notified when a new comment is posted: admittedly, because of the lack of threading, the new comment might not be in reference to what they posted.

Thanks, Suresh.

It sounds like what I really want is LiveJournal, managed by someone else a little less mercenary.

I think Vox is just a LiveJournal clone, run by the same company as LiveJournal as a branding thing — you know how one laundry detergent company will sell two or three differently packaged brands of laundry detergent as a way of capturing more market share? It was aimed at attracting an older audience than LJ itself was getting. But that was a buyout or two ago so I don't know whether it's still really run by the same company or worth paying attention to.

also, since community is important, it's going to be hard I imagine to move to another platform, since your community would have to move with you. In that way, friendfeed is trying something interesting, whereby you have a 'lifestream' that could merely be the feed from a blog, and others can interact with it via the friendfeed comment system (no threading though, but there is a kind of community feel to it)

Six Apart created Vox and bought LJ, but now LJ is owned by Sup and independent from Vox and other 6A properties.

Thanks. This post doesn't exactly make Vox look like a haven from ads, though. I wonder if Vox should be listed at this list of LJ clones — there's a strong resemblance in look and feel, but that's not enough for me to tell whether it's the same code.

I keep wanting to run the code in this post. i've got a string of related problems: once groups are determined, how can i evolve them over time based on new/removed connections, how can i organize them as a sequence/spectrum of related groups. it's basically the question of set size and overlap over a sequence of changes. Hmm here is an example:

http://tucc.us/11011110.xml

but then add events between associates that don't involve u as a party. see where i'm goin?

There's been a fair amount of research in problems like this, of drawing time-evolving graphs in a way that tries to avoid too much change in the layout from step to step, so that you can view it as a movie of how your graph changes. I'm not quite sure how that fits into the other problem of taking the huge graph of friend/unfriend events and determining which ones are near enough to the subject to be relevant, though.

i have been accused b4 of failing to do my homework. but yes that is the sort of movie i'm preparing. i want to assign letters to the groupings, and each event could include new lettered membership sets for the two parties, enabling maintence of groups in a client at low cost, and a great rendering clue. a spectrum ordering across the alphabet of sets with high associations would also be a great rendering clue. just a little more xml but dense with info. My solution for the step-to-step visual churn avoidance is determining an optimal final frame and favoring its general coloring and positioning when placing prior frames. We shall see...

In case it wasn't clear, BTW, that wasn't so much intended as an accusation of not having done homework, but rather a pointer to where you can find help with your homework.

ha ha no problem, i got it. about a month ago some academic told me my code was "unaware of the research" or something. i was a little annoyed by that guy. i want to get the IronPython dev environment running here and make your set code work. i'm still using a gnarly hack to find sets because buying computers has been easier than implementing new code. hmm i guess that's still true. i think if my sets code comes out clean it will be due to my pragmatic coding wits, and i think clever engineering is often as beautiful as clever maths. but complexity can deliver harsh punishments, too.

Grow up kid! You like a service and don't want to pay for it, either monetarily or a favor in return? How do you expect your service providers to survive?

In fact, Singaporean troll, I have been seriously thinking about upgrading to a paid service anyway, despite the fact that the money would go to a bunch of bloodsucking leeches, and despite the fact that I can accomplish many of the same things for free elsewhere. $20/year isn't that much to pay for the level of service I've been getting from LiveJournal, after all, and if I prefer LJ to use a user-supported financial model than an ad-supported financial model then it makes sense to use my money to encourage them in that direction.

The sort of hostile comment you just left, however, encourages me to do the opposite just to spite you. What are you expecting to achieve? I suppose the more appropriate response is to pretend you and trolls like you don't exist.

I only wish that it were possible to upgrade to a paid service that was exactly like LiveJournal except that its owners respected its users. The posting I linked to, in which they attempt to trick the users into thinking they're bringing back Basic when really what they're doing is turning existing Basic accounts into Plus-lite, is a case in point. It doesn't give me much confidence that they'll keep paid accounts ad-free (when have any of their other promises about keeping LJ ad-free been kept?) but I suppose upgrading to paid would give me the option of cancelling the paid service once they impose ads on that too. And it wouldn't be any harder to move somewhere else then than now.

Regardless of whether I pay for my account, here's why I think Basic should continue to exist, and be allowed for new users, in ad-free form: It builds the pipeline of new users that will eventually become paid users. If LJ requires new users to pay immediately or suffer ads, but other services continue to be both free-as-in-free-beer and ad-free, then why should anyone come to LJ instead of those other services? And I'd prefer the other users come to LJ because their presence makes it more of a community here.

$20 a year is not a lot of money. Sounds like a good decision. --Gabriel

I keep a copy of the source texts of all my posts, so if things here become bad, I can post a copy of them quickly elsewhere.

Yes, me too, using ljmigrate. But most of the time I think of it more as a backup and less of an escape option.