Open Advogato?

Advogato was never "closed" in the first place. You're free to change the code and contribute.

I think there are some people who don't wish to just modify the existing code and set up a new mod_virgule site, but who'd like to change the operations of the Advogato site itself.

Also, it's quite hard to get patches accepted by raph into the main CVS tree. Perhaps a semi-fork is in order?

Isn't mod_virgule already forked, at least a few times?

To me, it's advogato.org that matters anyway, not mod_virgule. I've submitted a patch or two to raph, without much effect. They'd be really useful to me while reading Advogato on my Palm and trying to reply to people...

what you do is you create an account and suck the ENTIRE content of advogato.org (including http://advogato.org/person/graph.dot and also http://advogato.org/rss/articles.xml) off the site.

repeatedly (every day? )

then you analyse it.

then you do your own thing.

then maybe you start cross-posting from your-own-thing to advogato.org but you would need to have people enter their advogato.org password into your-own-thing.org.

i've come across this you-can-always-fork one before: just because you can get the source code and run it on your own server, it doesn't automatically mean that you have the right to the USE of that code.

what you might want to consider is to extend advogato.org via the XMLRPC interface to have the concept of "<person>http://your-own-thing.org/person/myownthingaccount</person> instead of just myownthingaccount.

that's quite a dramatic, radical and very powerful change that requires an individual to explicitly enter into their advogato.org profile: "on advogato.org, i trust Certifications of type "Open Source Advocate" from site my-own-thing.org".

otherwise the whole trust thing just totally collapses as it is impossible to expect just any old XMLRPC-based site to have any form of security or trust or lack of corruption.

l.

After all why not to make it yourself?

I mean there is XML-RPC call support. So you could write a piece of software for own use which will do all you need. Just set it on your own server and it will post your GPGed mail to here and contain any JS for inserting HTML (up to contentEditable areas for Mozilla/IE).

Why fxn done it with MyAdvogato and you just grumble here?! ;)

And "the cutting edge" is not just "packed with feature". Read this article "Building Communities with Software".

In addition to code there a mailing list where patches can be sent. So if you want changes to advogato itself, that's a place to start.

The other side of the problem is that raph hasn't been spending much time on the site itself, so you either have to get his attention long enough to install an upgrade, or convince him you're a worthy delegate.

Actually, I think it's a testament to the technology how little maintenance advogato has needed over the years. I too would like to see some blog-inspired features of course, and there's long been a list of cool features waiting for someone to implement them, like diary threading.

So please do develop patches, or external utilities as Malx suggested. Show us the code.

My patch wasn't even that... If you go on a person's page, you see all those little "»" character beside each of their diary entries that are permalinks? Small and unobtrusive, right? My patch makes them appear in the recent logs too, that's all.

Well... while it's possible to (say) install MyAdvogato on some random server and tweak that, it's a bit problematic because it means that, even if raph suddenly gets lots of free time on his hands, there'll be no automatic way to synchronize improvements made to MyAdvogato and this main mod_virgule code. E.g. I can pound away at my PC for a few hours and add diary threading to MyAdvogato, but later on I'll still have to spend yet another few hours to add the same feature to mod_virgule... which sounds really silly (then again, this may be the only way).

Plus, I don't currently have a machine I can afford to turn on 24/7. :-(

Anyway, some time ago I did also send a patch for mod_virgule, which changes the "How interesting is XXX's diary" combo box to display the current rating instead of "--". I guess raph must've been really busy............

 It's true that there are many patches floating around for mod_virgule, the code behind this site.

 You can see some of them posted in peoples diaries, and others referenced upon the virgule-dev mailing list. (I even have a email my my password patch upon my website).

 There's a small list of patches and forks maintained by AdamShand, which is a good starting point at documenting the differences.

 I'd love to see a more current set of code here, but drew short of starting my own fork as I was working on the article editting features.

 Why? Because the code isn't the issue, sure it's in need of some patches, and sure many people have written changes - but what's really importent is the people who come here.

 If you're brave and adventurous you could setup a domain, host a current site and give commit access to a public CVS repository - but without visitors, without interesting articles, and without time you're site would be a pale imitation of the real site.

 I'm loathe to attempt to draw the crowd from here elsewhere - as I'm sure the other maintainers of forked code are. Otherwise it would have happened.

 (Hmmm somebody should write a voting module; then people could vote whether they'd move to another site if it was more featureful - if raph committed it ;)

 I think the best thing to do is to wait around and see if there is more time for the project from a committor, and if not, if the site disappears one day then would be the time to setup an alternative, rival, site.

 Thanks to raph for creating Advogato in the first place, and allowing me to interact with other people - people I would not otherwise have known.

 After all, what is the site without people?

Most of the existing forks (like mine, created to run robots.net) were made to run other sites. But I don't think any of those sites could be or even want to be replacements for Advogato. Nearly all the fork maintainers would like to see the official source maintained so we could pool our efforts and spend time improving one source tree instead of maintaining a bunch of dupicate forks. There was a valiant effort to merge the forks last year but it fizzled out before it got very far. The trouble seems to be that no one with CVS access wants to (or perhaps has time to) maintain the code and no one who wants to maintain the code can get CVS access.

Sorry guys - I really have most of the responsibility for the frustration people are feeling about getting their patches and improvements into Advogato. I've been focussed on work and family stuff a lot lately, and what time I have left over I'm trying to use for finishing my dissertation.

That said, there is hope. As sisob rightly points out, there's no compelling reason why the development of Advogato has to be so plodding. Basically, we just need a good volunteer or two to integrate patches into CVS and apply basic maintenance to the site. I'm pretty friendly, but I do expect people not to introduce gaping security holes and the like (this is, alas, all too easy to do in our rickety browser infrastructure).

lkcl's cross-certification ideas are interesting, but what this site needs is a mix of innovative trust work (which, of course, is one of the original motivations) and plain ol' good blog functionality.

I think the best place for this discussion to continue is the virgule-dev mailing list. See you there, and I'll do my best to be responsive!

I think it would greatly promote experimentation in trust metrics if something along the lines of lkcl's cross-certification ideas were to be implemented, with advogato being part of the network. This is non-trivial to get right, and because we would be establishing a protocol I think we should try to get the representation of the trust-metric right first time. I put some thoughts to electrons on a related theme a while back. By allowing assertions about cross-host account identity to be made, and together with a scheme for mapping certifications on one host to assertions to another, my scheme gives a fairly economical and flexible way of handling cross-certification. Criticisms? Worthwhile? Can you do better?

[i too would like to move this to virgule-dev, which i will "drop in on" once in a while. given that this is mainly non-technical discussions, here, this i believe is the best place]

the principle of trust metrics is the "top level seeds".

the idea behind using seeds in the first place is that the top-level seeds are "ABSOLUTELY" trusted (implicitly trusted) with respect to the Certifications.

floated several times has been the idea of using "local" trust to make "local" Certification decisions.

_you_ specify the seeds.

_you_ decide what to do with the Certification calculations based on _your_ choice of seeds.

nobody else gives a bugger-all care about what _you_ do with those Certifications: it's for your _own_ benefit.

as far as advogato.org is concerned, raph has decided, just like anyone else can, that _he_ is going to specify four seeds, raph, miguel, alan and federico, and what HE is going to do with those four seeds is make "posting" decisions on the site "advogato.org".

to cut a long story short, cross-host account identities need to be made useful in two stages:

1) add in user-selected "seeds" for user-selected "preferences" on viewing the site, advogato.org

2) add in an interface (XML-RPC?) to be able to add, view and remove Certifications remotely.

on a technical note, it is presumably possible to allow authentication (even as simple an authentication as adding user+password into the relevant XMLRPC function calls for example).

the rest is then ENTIRELY up to EXTERNAL users of the site to come up with weird and wonderful ways to use 1) and 2).

strictly speaking it isn't even necessary: register your site meta-advogato.org, write some code that grabs http://advogato.org/person/graph.dot, and off you go.

I'm skeptical of this trust metric thing and am not convinced that Advogato has been such a great testbed for them. That is a GOOD thing--I mean Advogato participants are generally well-behaved, post worthwhile stuff, and there are hardly any malicious users or trolls. But it means the high quality of Advogato content comes from its user community and there's no evidence that the trust metrics are helping it.

Is anyone running mod_virgule (or whatever) on sites with a big troll problem? For example, political discussion sites, pr0n sites (I guess they don't have much discussion though), or whatever.

The implementation as an Apache module written in C also makes mod_virgule difficult and risky to modify. Maybe it could be reimplemented as a CGI or FastCGI in a friendlier language (Python is my current favorite).

phr, i have the trust metrics code rewritten in python if you're interested. it's one third of the length of the c-code.

http://sf.net/projects/pymmetry.

i'd love to do a rewrite of mod_virgule as a mod_python or just a cgi: been meaning to for ages.

l.