![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
Entry tags:
Should I write an LJ interface to CommYou?
I think I need to take a survey of my friends. I'd appreciate it if you'd take a minute to fill out the following very brief poll -- it's likely to have a big impact on exactly how I shape CommYou in the very near future. (All of the long text that follows is by way of details and explanation: you can skip it and go straight to the poll if you want.) Feel free to point others to this, if you think they'd be interested.
Things are steaming along towards release 0.1, the very earliest and simplest alpha of CommYou. It's still ridiculously feature-light (it has less than a third of what I consider the necessary features), but almost getting real. My remaining steps are "Make The UI Suck Less" and "Break It Enough To Work With IE6". Once those are done, I need to write a proper build system, and then build the actual site. So I'm probably a few weeks from having the thing really up and running.
The question du jour is, what next? In particular: I've been planning, all along, to fully deploy this thing for Facebook before worrying about any other social networks, but I'm having second thoughts about that, especially since most of *my* friends are over here on LJ. I don't think LJ is quite such fertile ground as Facebook in the long run (partly because LJ already has adequate conversational tools, partly because it's simply a lot smaller), but I'm considering porting the thing to LJ simply so more of my friends are likely to try it out and participate in the development process. As quickly as possible, I'd like to be "eating my own dog food" (to use the software industry's standard slightly nasty metaphor) -- I would like to be having the discussions about CommYou's development and evolution *in* CommYou itself, so I want to get a bunch of interested people into it soon.
A quick review of the system itself: CommYou is superficially similar to LJ in some respects, but with a focus on conversation instead of blogging. The top posts aren't going to be quite as privileged as on LJ, and there is going to be a lot more tool support for the ensuing discussions. The biggest single difference, and the one that's visible upfront, is that the system tracks what you've already read and what's new to *you*, so you can more easily keep a whole bunch of conversations active at once. Also, the UI is much more interactive, and designed to make it a bit quicker to skim through what's going on. But there's lots more coming, ranging from tight IM integration to live online chatting to the ability to treat different conversations differently, depending on how important they are to you.
What CommYou is *not* is a social network: instead, it's designed to be a conversation system that works with and across social networks. The long-term game plan is that CommYou will integrate with *all* social networks, insofar as possible. Plenty of people are dealing with tracking who your communities are; I'm trying to make it easier to talk with them.
A LiveJournal adaptation of CommYou necessarily wouldn't be as tight as the one for Facebook -- LJ just doesn't have the hooks -- so it would basically be a separate site that knows about LJ and uses it. I believe you would sign into LJ via CommYou; CommYou would synchronize with LJ and import your friends and communities. You would be able to start and respond to conversations with your friends and in your communities inside of a CommYou webpage. The topics of those conversations could be posted into LJ in the appropriate places, but they would point to CommYou for the discussions themselves. I'd probably also support many of the LJ clones (GreatestJournal, DeadJournal, etc) quickly, simply because it would be easy to do. Within some limits, CommYou would merge your identity (if you wanted) across all social networks it knows about: initially, this probably means that you could share conversations with friends in both LJ and Facebook.
So, given all that, the question is: would you use it?
[Poll #1156733]
Things are steaming along towards release 0.1, the very earliest and simplest alpha of CommYou. It's still ridiculously feature-light (it has less than a third of what I consider the necessary features), but almost getting real. My remaining steps are "Make The UI Suck Less" and "Break It Enough To Work With IE6". Once those are done, I need to write a proper build system, and then build the actual site. So I'm probably a few weeks from having the thing really up and running.
The question du jour is, what next? In particular: I've been planning, all along, to fully deploy this thing for Facebook before worrying about any other social networks, but I'm having second thoughts about that, especially since most of *my* friends are over here on LJ. I don't think LJ is quite such fertile ground as Facebook in the long run (partly because LJ already has adequate conversational tools, partly because it's simply a lot smaller), but I'm considering porting the thing to LJ simply so more of my friends are likely to try it out and participate in the development process. As quickly as possible, I'd like to be "eating my own dog food" (to use the software industry's standard slightly nasty metaphor) -- I would like to be having the discussions about CommYou's development and evolution *in* CommYou itself, so I want to get a bunch of interested people into it soon.
A quick review of the system itself: CommYou is superficially similar to LJ in some respects, but with a focus on conversation instead of blogging. The top posts aren't going to be quite as privileged as on LJ, and there is going to be a lot more tool support for the ensuing discussions. The biggest single difference, and the one that's visible upfront, is that the system tracks what you've already read and what's new to *you*, so you can more easily keep a whole bunch of conversations active at once. Also, the UI is much more interactive, and designed to make it a bit quicker to skim through what's going on. But there's lots more coming, ranging from tight IM integration to live online chatting to the ability to treat different conversations differently, depending on how important they are to you.
What CommYou is *not* is a social network: instead, it's designed to be a conversation system that works with and across social networks. The long-term game plan is that CommYou will integrate with *all* social networks, insofar as possible. Plenty of people are dealing with tracking who your communities are; I'm trying to make it easier to talk with them.
A LiveJournal adaptation of CommYou necessarily wouldn't be as tight as the one for Facebook -- LJ just doesn't have the hooks -- so it would basically be a separate site that knows about LJ and uses it. I believe you would sign into LJ via CommYou; CommYou would synchronize with LJ and import your friends and communities. You would be able to start and respond to conversations with your friends and in your communities inside of a CommYou webpage. The topics of those conversations could be posted into LJ in the appropriate places, but they would point to CommYou for the discussions themselves. I'd probably also support many of the LJ clones (GreatestJournal, DeadJournal, etc) quickly, simply because it would be easy to do. Within some limits, CommYou would merge your identity (if you wanted) across all social networks it knows about: initially, this probably means that you could share conversations with friends in both LJ and Facebook.
So, given all that, the question is: would you use it?
[Poll #1156733]
no subject
I don't have time to deal with Facebook; I have vague, ill-informed prejudices that it is largely about expressing teenage angst, and that my friends are much more likely to pick LJ. I also have it juxtaposed with MySpace, probably unfairly.
Two technical question/suggestions: do you have stories about authentication via OpenID, or interoperation with Ning?
no subject
Oh, absolutely, but that's intentional. This poll is very specifically about whether there are enough LJ-only users who are interested in playing with this for it to be worth my while to port it. It's not a small effort -- certainly it would divert my attention for at least a couple of weeks -- so I only want to do it if there seems to be sufficient interest.
I don't have time to deal with Facebook; I have vague, ill-informed prejudices that it is largely about expressing teenage angst, and that my friends are much more likely to pick LJ. I also have it juxtaposed with MySpace, probably unfairly.
Honestly, I've heard the same criticism leveled at LJ -- I think people say it about all social networks. In practice, I think the angst in general is actually a bit less *visible* on Facebook, precisely because its communication tools are such crap. As for the difference in the average age, that's a bit hard to say. Last I checked, LJ's public demographics looked fairly similar to Facebook's (that is, much younger than average for our crowd), but I don't have much faith in those. I'm a bit cynical about Facebook, but only because it has been so clueless on the communications front -- other than that, I find it a reasonable tool, and innovative (if rather horribly quirky) in a number of respects.
(And as for MySpace, it's a fairly apples-to-oranges comparison. Facebook and MySpace are probably about as similar as Facebook and LJ. They all have very different styles and approaches to social networking. That said, I don't know MySpace *nearly* as well as Facebook.)
All that said, unfair biases are okay. I'm looking right now for honest answers about how many interested LJ users would be unlikely to give it a try if it started out Facebook-centric. The reasons why not aren't terribly relevant -- frankly, getting a statistical sense of the bias, whatever its reasons, is the most valuable thing to me right now. There might or might not be a later time for me to try and cajole people into getting FB accounts to use CommYou, but this isn't that time.
[cont'd]
no subject
Yes, and sort of.
I've been assuming that OpenID would be an important part of the equation for a long time (indeed, since before the beginning of the project), but it's not enough in and of itself. I'd prefer to use OpenID for authentication, but I also need to be able to fetch the social network information, which isn't inherently part of OpenID.
As a general principle, the less sensitive user data I have to be handling, the better. I have my own fish to fry -- the conversation data -- and that's a big enough task for me. So my preference is to use, eg, something like Paypal for any monetary transactions (so I don't have to handle credit card data) and OpenID for authentication (so I don't have to store user names and passwords). How closely I can approach those ideals remains to be seen.
(Indeed, I'm in a very interesting situation. CommYou, ideally, doesn't actually have its own logins: it relies *entirely* on other services for its authentication, and its accounts are stated in terms of those in the DB. I don't know many other services that can say the same yet, although I suspect it's the coming wave.)
I was actually thinking about exactly that this morning. The interesting question for my purposes isn't whether *I* accept OpenID -- it's whether LJ's own APIs do. My preferred architecture would be to authenticate via OpenID, thereby getting an authenticated token that I could pass into LJ's APIs. I'm not at all sure yet whether the APIs talk to the OpenID side, though -- that's a topic for research. Once I decide to get serious about LJ integration, then I need to understand how that all hangs together.
As for Ning, I'm not worrying about that specifically, but it's part of the much larger and more important story of covering OpenSocial. I've been starting with Facebook because it's the more mature (if idiosyncratic) platform, but I'm assuming that OpenSocial will become crucial fairly soon, since that covers more or less everybody *except* Facebook. So that's a "when", not an "if".
It's also possible that some social platforms would benefit from additional customization, beyond what OpenSocial provides. I'll take those on a case-by-case basis. The high priority is going to be covering the main common case.
This may also play into the LJ equation, of course. So far, I don't have any clear evidence about LJ's support for OpenSocial, aside from some very airy claims. If LJ does turn out to support it soon, then that would probably lead me towards simply tackling the OpenSocial port soon, and treating LJ as a special case...
no subject
It is a decent way to get a contact to old friends, though.
no subject
It will keep you from making subconscious design decisions that work one place, and not another.
Make one of your two platforms the "most crippled place that it can possibly work". This means it is truly portable, and also: it forces you to have an architecture that lets you extend the product to provide more features in a more mature environment.
no subject
But yes -- this was one of the underlying motivations behind this survey. In particular, dealing with another platform would force me to get my schema straight, and that's the really high priority. UI and mid-level details can be adjusted later, but changing the DB in big ways is a major pain in the ass.
So only half the question for me is "do I do a port now?" -- the answer is preferably yes, but that's *far* from decided. The other question is whether that port should be to LJ (because it's where I know a lot more people who might poke at it in the short term) or to OpenSocial (which is a much more important platform in the long run)...
no subject
It might be enough to support OpenID.
no subject
no subject
no subject
(One of the ideas we talked about at Zingdom, but never got around to doing anything with, was analyzing the social and adoption graphs, and giving some sort of bennies for the main nexi in that, to encourage them. I still don't know exactly what that *means*, but it lurks in the back of my mind, and I'll probably at least loosely track the adoption graph...)
no subject
Within some limits, CommYou would merge your identity (if you wanted) across all social networks it knows about: initially, this probably means that you could share conversations with friends in both LJ and Facebook.
...if you provide an effective, incremental migration path off of LJ, you will get more users than you know what to do with the next time something like this happens. (And it almost certainly will; it has, historically, and the linked-to attitude of the new owners doesn't make it look like it's any less likely to happen again.)
I'm guessing you're not at a point where you want to drive hard for that - too early, too alpha.
But it's something to keep in mind as a future means of getting people onto the service.
no subject
Migration
That's what I thought of, too. If CommYou had its own backend based on the LJ code, it could make it feasible to move from one LJoid to the other and still have all the CommYou settings.
However, I would have to think hard about giving CommYou my password so it could access LJ—not as hard as if it were run by a stranger, but still.
Re: Migration
Not gonna happen -- CommYou is going to wind up *way* more sophisticated than LJ, I suspect, and I don't want to limit myself with its assumptions. Indeed, getting my schema right is one of my biggest challenges. I'll be lifting ideas from the LJ backend (for instance, the IT plan has integration with memcached likely in the medium term), but I need to go well beyond it.
However, I would have to think hard about giving CommYou my password so it could access LJ—not as hard as if it were run by a stranger, but still.
Well, unless I miss my guess I'll be able to avoid actually storing peoples' passwords anywhere. If things go *really* well, I can just use OpenID or OpenSocial for authentication, so you don't have to trust CommYou with it at all. If that doesn't work, I'm hoping to only have the password in a transient way: use it to log in, get a session token, and only store that in the CommYou DB.
See my response to
Re: Migration
Oh, I didn't mean the CommYou codebase would integrate LJ code; I just meant that the CommYou corporation would run an LJoid server alongside your own codebase. You'd need code to talk to it, but that'd be the same code as talking to LJ. Then LJ users whose conversations had mostly moved to CommYou could get off of LJ proper and keep the same conversations.
I know that's hard; IIRC, SixApart canned the "export my account" functionality years ago. But it wouldn't be something for the short run anyway; you'd need a critical mass before LJ users would be willing to make the switch.
That would be good.
That'd be better than long-term storage, but you'd still have a pretty large job of persuasion to do—that's essentially the model with debit card POS terminals, and TJX showed us that we can't rely on it all the time.
One question that'll make a difference: if you've got a session token to my account, and I change my password, is your token invalidated? If so, then it's easier to trust you with that token, because I can withdraw it unilaterally.
Re: Migration
Hmm. Could be done, but I don't know if it's worth it. If I was going to go that route, it might make more sense to simply adapt the data into CommYou proper. That would probably *mostly* work: there are some mismatches in the model (for instance, I'm not sure my conversations would look good as deeply threaded as LJ tends to get), but the guts would mostly merge in okay, especially if I was smart about it.
I'm a little skeptical that anyone's going to care enough, honestly. But it's an interesting idea to salt away in the back of my brain for later.
That'd be better than long-term storage, but you'd still have a pretty large job of persuasion to do—that's essentially the model with debit card POS terminals, and TJX showed us that we can't rely on it all the time.
Oh, sure. But that's why I'm fairly likely to at least partially open my code, even if I don't go the full open-source route: so that it's not just my word that people are relying on. (And so that the white-hat community has the ability to point out problems before they become disasters.)
In general, though, there's no substitute for reputation. If I can manage to develop a reputation as a trustworthy system, I'm more likely to be trusted. That comes only with care and time. That applies quite generally: I expect to have to earn peoples' trust to hold the conversation data properly, to begin with.
One question that'll make a difference: if you've got a session token to my account, and I change my password, is your token invalidated? If so, then it's easier to trust you with that token, because I can withdraw it unilaterally.
No clue, although I doubt it. It'll depend on the details of the API, which I haven't looked at enough to know...
no subject
Hadn't seen that. I'm *astonished* by the claim that base accounts have only been accounting for 10% of all new registrations. If that's true, then I somewhat change my mind about the sensibility of the decision: it was grabbing the tiger by the tail for no particularly good reason. I had assumed that base accounts were still making up 70% or so of the memberships. (OTOH, if it's true, then the decision to simplify the choices is more reasonable, and it probably means that they are right that this will all blow over.)
The interview in general -- really, I'm mostly taken aback by the fact that he's so forthright about it. Definitely pretty cynical, but also probably mostly correct. The main thing he's missed is the social dynamic of social networks, and the fact that cynicism at the top can be gradually corrosive. So I think he's correct that the current firestorm will die down, but he's missing the fact that this is likely to gradually eat into the business.
But it's something to keep in mind as a future means of getting people onto the service.
Oh, I am *well* aware of that. By decoupling conversation data from social networks, CommYou potentially makes it significantly easier to switch networks, or to support several of them. Indeed, it's the biggest reason why I expect that, if I'm successful, I'm going to gradually make the big players a tad cranky.
But I don't want to get my hopes up too far on that score. CommYou is still going to be another commercial entity, not a homebrew hobbyist site, and some of the people who jump ship from LJ will be understandably suspicious of it. I'm hoping to do a better job than SA and SUP have -- in large part, by being more upfront about the social contract and working out the expectations better in advance -- and I like to hope some people will decide that CommYou provides just this sort of migration path. But I'm not going to rely on that too heavily.
And yes, it's *much* too early right now. I'm only just past the "prototype" stage -- it's usable, but best described as being at a Google first release level of functionality. (That is, preposterously stripped to its core.) That's why I'm mostly looking for friends to kick the tires: I don't think it's nearly good enough to open to even a private beta yet. And it's going to be a fair while before it is feature-rich enough for people to consider it a viable alternative to LJ, at least for the blogging-centric use cases...
no subject
Question: would this be all one, er, thing? So someone with a Facebook account and someone with an LJ account could talk via CommYou, not have the FB-CommYou and the LJ-ComYou separate?
no subject
Yaas. They're not as immediately relevant to me (since what I'm mainly looking for right now is my friends, who are mainly on LJ), but once the port is done, there probably won't be any reason *not* to support at least the larger spinoffs.
Question: would this be all one, er, thing? So someone with a Facebook account and someone with an LJ account could talk via CommYou, not have the FB-CommYou and the LJ-ComYou separate?
Ideally, yes. There are some technical challenges (and even more, legal challenges -- doing it without violating Facebook's Terms of Service will be tricky), but that's the objective. Indeed, one of the major complications is likely to be the "in-between" states: allowing you to merge you LJ and FB identities on CU, but allowing you to post, eg, something to just your LJ friends. I suspect that will be a requested story, so I'll need to take it into account...
no subject
Is "Post something in a CommYou conversation which takes advantage of multiple distribution channels [eg, LJ and FB], but doesn't show up twice for someone else who's on CommYou pulling from both LJ and FB" covered already?
no subject
no subject
Try versus adopt
If you read my entry, you know I think this is a _big_ idea that has the potential to grab a strong niche.
Besides, I'd be curious to see how CommYou interacts with LJs ads...
Re: Try versus adopt
Oh, absolutely -- I wouldn't expect anyone to commit to really using it at this point. Hell, *I* don't think it's good enough to make such a commitment yet. But part of the goal here (indeed, the main goal) is to get enough people playing around with it so that we can collectively figure out what it needs in order to get it to the "good enough" point.
If you read my entry, you know I think this is a _big_ idea that has the potential to grab a strong niche.
Haven't gotten to it yet (Mr. Perpetually Behind strikes again), but thanks. This is really eating my brain right now, mostly because I'm pretty sure that I'm on to something.
Besides, I'd be curious to see how CommYou interacts with LJs ads...
Mostly not relevant, actually. Consider it this way: CommYou is going to be its own user interface. (Necessarily: the kind of in-page integration I'm doing with Facebook isn't possible in LJ.) I'll probably post pointers into LJ that go to the CommYou conversation, but you'll actually read the conversation inside CommYou. So those pointers are really the only place that LJ *can* inject ads.
Re: Try versus adopt
You'll need to corral some number of our friends who use the ad-supported level of LJ, for their feedback, since you and I are Permanent and don't see the ads. (Or create a test account)
Re: Try versus adopt
no subject
no subject
no subject
no subject
No -- it's precisely *not* that. Actually, it probably can't be: LJ's APIs for accessing conversation data are pretty crappy, AFAIK. (For all that it talks about being very open, it's often fairly hard to use externally.) At best, there might be something akin to an RSS-feed view of LJ, using its posts as conversation starters, but it's unlikely to be feasible to integrate LJ's actual threads into CommYou.
This is a separate collection of conversations, with the same friends that you have in LJ. Same login; same communities; same friends; but separate conversations and a somewhat different mode of interacting with them. There will be a bit of cross-talk possible, but it's necessarily pretty limited.
Mind, this is exactly why Facebook was the first target, and why my plans have generally been to do LJ very late, if ever. Facebook doesn't really *have* conversations in any meaningful sense: it's a much purer and more typical social network. LJ is weird and complicated, because it has conflated the social networking with the blogging and conversation, so it's harder to build on top of. (And unfortunately, this tends to result in LJ-only users expecting social networks to be far more than they typically are, which often causes confusion when I'm trying to explain CommYou to them.)
So basically, CommYou would mainly use LJ as another social network -- a graph that its conversations get laid upon. Any additional interaction beyond that is largely gravy...
no subject
no subject
I do have a Facebook account. I made one to try it out and see what the buzz was about.
I find it TOTALLY USELESS. It is the modern equivient of the early 90's "This is my little corner of the web" personal pages. Naive, eager and childish. I have no idea why it is popular. Even the limited ability to hook you back up to people you went to school with is better served in other locations. If I wanted to be found again, I would go do the seeking myself. I find the little programs associated with it "poke someone!" "get into a pillowfight!" a waste of anyone's time who is over the age of about 15 and not trying to get the girl next door in bed.
This is so irritating to me because the potential power of Facebook is being squandered.
If CommYou can actually make Facebook useful then I would use my Facebook account to play around with it (and I would make you cookies to thank you for putting the "ace" in Facebook).
no subject
Ayep. That particular observation, about six or seven months ago, was actually the genesis of this project. My standard take on it is that there are around 50 million people with Facebook accounts, and 3/4 of them have no damned idea why. Indeed, FB themselves are well aware of it, and scared silly AFAICT.
While I may not ever commit it to paper, I've occasionally toyed with flyering the college campuses with something that reads, "Yeah, we know -- Facebook kinda sucks. Want to be able to actually *do* something with it? Try going to this URL, and check it out." I might yet try that, if I decide that I need a little guerrilla marketing...
no subject
no subject
(But I'm very partial to butterscotch...)
no subject
I have a whole mob of people I know from political blogs who are now on Facebook because there was a project about a year ago to try to build something useful on FB for political organizing. I could probably get some of them on board for CommYou at some stage; let me know when you'd like me to try that. (I'm thinking 0.1 probably isn't the time.)
no subject
Probably true, but maybe for 0.2. (Which I'm calling either "alpha 2" or "beta 1" -- haven't decided. It's the point at which the features are starting to fill out a little, and the Facebook integration is getting decent.)
no subject
You might want to take a look at Ravelry (ravelry.com). Their forum system is doing this very thing.
no subject
Thanks for the pointer!
no subject
no subject
So it's sort of like a large number of chat rooms. In that respect, it's really a lot like LJ, just with the emphasis shifted slightly from the top post to the resulting discussion...
no subject
no subject
no subject
no subject