Google Group (was: Why not forget about AltMe)

great ;) although something less "beta" might be preferrable, but as long as it behaves like a normal mailing list (and google groups does this) it's fine for me ;) -- Best regards, Andreas

What degree of spam protection is there on this group? The more useful usenet groups are moderated. The unmoderated ones can end up like this (admittedly, an extreme example): http://groups-beta.google.com/group/comp.software.year-2000 Sunanda.

Currently all shields down. We can switch on member-moderation and message-moderation. then they need approval.

-- -Volker Any problem in computer science can be solved with another layer of indirection. But that usually will create another problem. David Wheeler

Thanks. I just joined the group, posted a couple of messages, and checked the available security levels. They seem primitive and the third -- which is the one you refer to, means no message will be indexed by goole for public consumption. That makes the GoggleGroup almost as hidden as an Altme world. How would new people find it as ask to join? From the GoogleGroups "Create a Group" page: Access Level [ ] Public - Anyone can read the archives. Anyone can join, but only members can post messages. [ ] Announcement-only - Anyone can read the archives. Anyone can join, but only moderators can post messages. [ ] Restricted - People must be invited to join the group and post or read messages. Your group and its archives do not appear in public Google search results or the directory. Sunanda.

The checkboxes as moderator look a bit different. This should be possible: Who can become a member of your group? -> "Membership is moderated" Who can read the group archive? -> "Anyone can read the archive" Who can post messages to your group? -> "Members only" Should messages be moderated? -> "Messages are moderated" From help: "How do I approve pending posts in my moderated group?" http://groups-beta.google.com/support/bin/answer.py?answer=7937&topic=110

-- -Volker Any problem in computer science can be solved with another layer of indirection. But that usually will create another problem. David Wheeler

Thanks again. Definitely more flexible from the inside. The biggest remaining problem I see is that GoogleGroups postings are only returned in Google search results if someone has explicitly clicked the "groups" tab. Be honest, how many of us do that regularly when looking for information? usenet is 95% a waste of time, so it's almost not worth looking at. And, related to that only Google has a full index of usenet -- other search engines do not. If Google powers 60% of all searches (and assuming that number stays as high) then 40% of all searches will fail to find the group at all. And only a tiny minority of all other searches (those where someone clicks the Google Groups link) have a chance of finding the group. That's almost as invisible/exclusive as an Altme World. A big step backwards from the ML, Sunanda.

Why not writing a small script that publish mails on rebol.org ? So, we have two archives : on rebol.org, and on google... The rebol.org is indexed by google, and is retured for rebol-features requests. If this script is easy to write, or exists (?), I don't see where is the problem :-D --Charles.

We'd have to seek permission to republish. GoogleGroups are claiming rights over the postings (which, they claim, are licensed to them)....Item 4 here: http://www.google.com/googlegroups/terms_of_service.html Not impossible -- I'm sure they are still reasonable people. Does someone in favour of using GG want to check this out? Sunanda.

I think it should be really easy to provide a (second, third, fourth, whatever?) public, web-accesible archive on another site, like e.g. rebol.org. -- Best regards, Andreas

Glad you have this opinion! :) Carl mentioned that too ( i admit i obverlooked it). And everyone on altme suggested a second web-list. Like the one on rebol.org for old ml. So google would index it. And as you are the expert for such things.. would you? Would be great! :) (Oh, seeing some other mails here suggesting it, hi all! :)

-- -Volker Any problem in computer science can be solved with another layer of indirection. But that usually will create another problem. David Wheeler

Following some altme messages and some private emails, we've got a nice Community collaborative exercise running here..... ** Andreas is checking with Google whether we can republish posts from their GoogleGroup (their TOS can be read optimistically, but are not completely clear). ** Volker has updated the GoogleGroups Join message to notify new members that their posts may be archived elsewhere on the web ** Volker is also looking at extracting posts from the GoogleGroup to feed them to REBOL.org......Not an easy task given that GoogleGroups offer no POP-style API. ** If he can do that, I can add them to the Mailing List archive at REBOL.org using existing code (though it may need some strengthening for spam filtering -- GoogleGroups is a lot more vulnerable to sudden spamming than Ecartis is. I'm still thinking about that). In many ways, I hope this will be a trial run for whatever replaces the existing mailing list (if anything does -- it's not dead; it just developed a bug a month or two back). But, whether GoogleGroups becomes the main discussion forum, or it becomes an important secondary hangout, or we move to another technology altogether, we should end up with the ability to have a single integrated discussion forum archive at REBOLorg. And it's good to see the Community leaping into action in such an enthusiastic way, Sunanda.

While I can't see a problem with an archive on rebol.org and Google's Terms of Use (in my opinion Item 4 is only applicable to Google-owned content like the FAQ, whereas Item 7 applies to postings), I agree with Sunanda that an explicit statement/answer from Google is desirable. I've sent an email to groups-support on 2005-01-06:

No reply yet. -- Best regards, Andreas

Sadly, for full resilience, it's not as simple as that. Even the current connection between REBOL.org and the ML is more layered than that. As Carl (and others) have pointed out, the problem with email servers is that they don't always deliver emails. If REBOL.org was just a member of the ML, it would be failing to receive 15% of emails -- just as many of the rest of us are. (At one point, REBOL.org was losing those messages). To be dependant solely on an email-based feed from GoogleGroups to REBOL.org -- would be to set up the same single point of failure that we used to have. More resilient solutions are (or should be) available. It still needs some thought. And I am still not convinced that GoogleGroups is the best starting point. But the principles involved will be similar, not matter where the initial publication takes place, so what Volker is working on is a good pilot project. Sunanda.