Major correction to your footnote two: Virtually all-- just a select five or so (MSN, Yahoo, AIM, ICQ, and GaduGadu) aren't compliant.z-man wrote:Ah, joy. Two out of three predictions correct. Minor correction to my footnote two: Jabber or Google Talk.
Internal community chat
-
- Dr Z Level
- Posts: 2246
- Joined: Sun Mar 20, 2005 4:03 pm
- Location: IM: luke@dashjr.org
You're not a mathematician my chance? For them, "almost all" means "all except for a finite set". Which is, of course, always true if the whole set is finite. That the users of the excluded set of IM programs make up like (*) 99% of the total IM users is completely irrelevant to you.
This rant is not about whether XMPP is a good protocol to support, by the way. It would be perfect if Armachatron would use it, the last thing we need is another IM protocol around. It simply is about you completely ignoring facts even though they have been pointed out to you previously. Repeatedly, if I may add. Malevolent natures would say you are deriberately spreading misinformation. And since you're a project developer, that can cast a bad light on the whole project. So please stop. Every time, I have, to correct you, do research which you never bother to do. And I don't enjoy that.
On topic: why not give the Armachatron server an IRC interface? Map every channel to an IRC channel.
(*) Yes, I partly made that one up. The only census I could find gave the combined non-XMPP messengers 126% market share because it didn't eliminate people who use MSN at work and AIM at home and was from 2004. However, the only client not in your list that made it into that particular poll was Trillian (supporing Jabber, I assume) with 1%. Soure: http://www.pewinternet.org/pdfs/PIP_Ins ... Report.pdf
This rant is not about whether XMPP is a good protocol to support, by the way. It would be perfect if Armachatron would use it, the last thing we need is another IM protocol around. It simply is about you completely ignoring facts even though they have been pointed out to you previously. Repeatedly, if I may add. Malevolent natures would say you are deriberately spreading misinformation. And since you're a project developer, that can cast a bad light on the whole project. So please stop. Every time, I have, to correct you, do research which you never bother to do. And I don't enjoy that.
On topic: why not give the Armachatron server an IRC interface? Map every channel to an IRC channel.
(*) Yes, I partly made that one up. The only census I could find gave the combined non-XMPP messengers 126% market share because it didn't eliminate people who use MSN at work and AIM at home and was from 2004. However, the only client not in your list that made it into that particular poll was Trillian (supporing Jabber, I assume) with 1%. Soure: http://www.pewinternet.org/pdfs/PIP_Ins ... Report.pdf
Anoter easy way to access our IRC channel is provided directly by freenode:
http://java.freenode.net/index.php?nick ... armagetron
http://java.freenode.net/index.php?nick ... armagetron
-
- Dr Z Level
- Posts: 2246
- Joined: Sun Mar 20, 2005 4:03 pm
- Location: IM: luke@dashjr.org
The topic was IM clients, not IM users.z-man wrote:You're not a mathematician my chance? For them, "almost all" means "all except for a finite set". Which is, of course, always true if the whole set is finite. That the users of the excluded set of IM programs make up like (*) 99% of the total IM users is completely irrelevant to you.

While 99% might use non-standards-compliant IM services, there is also a large number of people using standards-compliant IM as well. Before Google Talk was released, there were more people using compliant services than were using ICQ (which is now part of the AIM network).
What facts am I ignoring? I've never claimed the standards were followed by a majority of people, just by a majority of clients and that it solves many of our tasks.z-man wrote:This rant is not about whether XMPP is a good protocol to support, by the way. It would be perfect if Armachatron would use it, the last thing we need is another IM protocol around. It simply is about you completely ignoring facts even though they have been pointed out to you previously. Repeatedly, if I may add. Malevolent natures would say you are deriberately spreading misinformation. And since you're a project developer, that can cast a bad light on the whole project. So please stop. Every time, I have, to correct you, do research which you never bother to do. And I don't enjoy that.
No, the topic here is ady developing an own IM. Slightly offtopic was already the relevance of the XMPP protocol for practical purposes. And for that, the number of supporting clients is completely irrelevant, the number of actual users counts. I bet you would also advise a company to develop software for Linux only, not Windows; Linux is more relevant because there are more Linux distributions than Windows versions.Luke-Jr wrote: The topic was IM clients, not IM users.
First, show me the numbers. Second, nobody suggested using the ICQ protocol. Third, ICQ, according to the poll I linked to, has the smallest user base of the big four proprietary systems. So now you're recommending to develop for Linux solely because it has more users than Zeta.Luke-Jr wrote: Before Google Talk was released, there were more people using compliant services than were using ICQ (which is now part of the AIM network).
Just because what you say is not actually wrong doesn't make it not misinformation.
-
- Dr Z Level
- Posts: 2246
- Joined: Sun Mar 20, 2005 4:03 pm
- Location: IM: luke@dashjr.org
And the closed analogy for QT in our IM case would be using a multiprotocol library, should such a thing exist, to link Armachatron with as many IM users as possible. Should such a thing exist, of course. But what you wouldn't do is pick a protocol of admittedly high moral value (XMPP) over a protocol of higher practical value.
I'd say in this case the highest practical value of any single protocol would come from server side IRC integration via a bridge. If we suppose the app is any good, the Windows guys will be happy with it. The Linux guys are, overall, happy with the IRC side. I'm not sure about the habits of the average Mac user, but I suppose is at least an acceptable way for them.
I'd say in this case the highest practical value of any single protocol would come from server side IRC integration via a bridge. If we suppose the app is any good, the Windows guys will be happy with it. The Linux guys are, overall, happy with the IRC side. I'm not sure about the habits of the average Mac user, but I suppose is at least an acceptable way for them.
-
- Dr Z Level
- Posts: 2246
- Joined: Sun Mar 20, 2005 4:03 pm
- Location: IM: luke@dashjr.org
Note the XMPP transport/gateway API that permits interfacing to non-standard protocols.z-man wrote:And the closed analogy for QT in our IM case would be using a multiprotocol library, should such a thing exist, to link Armachatron with as many IM users as possible. Should such a thing exist, of course. But what you wouldn't do is pick a protocol of admittedly high moral value (XMPP) over a protocol of higher practical value.
There are already many XMPP to IRC gateways, and FreeNode was at one point (before lilo died) looking into native XMPP support.z-man wrote:I'd say in this case the highest practical value of any single protocol would come from server side IRC integration via a bridge. If we suppose the app is any good, the Windows guys will be happy with it. The Linux guys are, overall, happy with the IRC side. I'm not sure about the habits of the average Mac user, but I suppose is at least an acceptable way for them.
I'm well prepared to find out that "many" means "two" and that those are stuck in a very basic state of development
But I'd take any solution. And of course, it remains to be seen whether it would be better to go Armachatron - XMPP - IRC or Armachatron - IRC - XMPP as each conversion is likely to cost features. I'd say it depends on the way Armachatron handles things.

-
- Dr Z Level
- Posts: 2246
- Joined: Sun Mar 20, 2005 4:03 pm
- Location: IM: luke@dashjr.org
About 25 or so by my counting.z-man wrote:I'm well prepared to find out that "many" means "two"
It seems to support kicking, but not opping or banning, or any of the other few features IRC has.z-man wrote: and that those are stuck in a very basic state of development
Of course, only guru3 has ops on #armagetron anyway, and he uses a regular IRC client already.
I don't see any possible way to make a gateway from IRC to anything, based on the protocol and how IRC was designed. IRC was designed to be a single (distributed) server, it can't handle the concept of others existing.z-man wrote:But I'd take any solution. And of course, it remains to be seen whether it would be better to go Armachatron - XMPP - IRC or Armachatron - IRC - XMPP as each conversion is likely to cost features. I'd say it depends on the way Armachatron handles things.
Could you elaborate on the gateway thing? I'd say a gateway should ideally be symmetric, it doesn't matter whether it is an XMPP-IRC or IRC-XMPP gateway. However it is called, I'd expect a gateway to transport everything it can from each side to the other. Do you mean that you can make an IRC channel look like an XMPP room, but not the other way round?
We'll probably need your list of 25 or so gateways later. By the way, Ady, how are things progressing? Have you hit the "it's 80% finished, but strangely, 80% of the work is still to do" wall? Don't worry, we all have at one point or another
We'll probably need your list of 25 or so gateways later. By the way, Ady, how are things progressing? Have you hit the "it's 80% finished, but strangely, 80% of the work is still to do" wall? Don't worry, we all have at one point or another

-
- Dr Z Level
- Posts: 2246
- Joined: Sun Mar 20, 2005 4:03 pm
- Location: IM: luke@dashjr.org
Right, because of the nature of how IRC is. To do a "real" gateway, you need an authorized server to communicate using IRC's s2s protocol (thus making it not a gateway, but a native XMPP interface). Without that, best you can do is implement c2s for each session.z-man wrote:Could you elaborate on the gateway thing? I'd say a gateway should ideally be symmetric, it doesn't matter whether it is an XMPP-IRC or IRC-XMPP gateway. However it is called, I'd expect a gateway to transport everything it can from each side to the other. Do you mean that you can make an IRC channel look like an XMPP room, but not the other way round?
That's on jabber.org's public server list.z-man wrote:We'll probably need your list of 25 or so gateways later.
-
- Dr Z Level
- Posts: 2246
- Joined: Sun Mar 20, 2005 4:03 pm
- Location: IM: luke@dashjr.org
Oh, nope. There's not much competition in the implementation of a gateway, due to the nature of their design (exacted design specs, nothing based on opinions or values).z-man wrote:Ah, so you mean "machines running gateway software" with that countI thought you meant different implementations.
A native XMPP interface to an IRC s2s protocol would be a good thing tho... I might consider writing one if I ever get time.