Uh.... AIM servers (and ICQ too now) listen on every port, so you can't block their AIM and ICQ access based on port alone. And actually the AIM client's network setup will bounce through random ports searching for one that works, so they don't even have to try that hard to get around it. I don't know about MSN and Yahoo for certain but I think they only listen on one port. But if you think you're blocking AIM, you're wrong.
Use it for compatibility with Microsoft, not to make Linux strictly better. Note that I didn't say we'd steal their scheduler, or paging algorithms, or anything like that. File system implementations and drivers are the kinds of things where all the systems that use them have to function nearly identically. I personally feel that ReiserFS, ext3, XFS, JFS, etc. etc. are better than NTFS, and that drivers where we have the specs are better than their windows equivalents. It's the drivers and file systems that we don't have the specs for that Linux isn't always so hot on.
I don't feel that's ironic or hypocritical at all.
What would I do with it? Probably just oo and ah. What do I expect people would do with it? Well, let's see.... How about getting NTFS in Linux working well? Depending on how much they release I'm sure there are some other drivers that have been reverse-engineered that might get clued into a few things. If they also have to release Office source (not that I expect they'd have to, since Office was never part of the whole comingling thing), get all those open source projects to properly read and write Office formats.
The worst thing that can happen for Microsoft is for users to have viable alternatives. Compatibility increases the chances of that. Source improves compatibility.
Hi. I guess I should answer, seeng as how, as the AC said, I'm doing 99% of the work. Sigh, and I had moderator points and was looking to use them, too.
Gaim has actually transferred hands two or three times (depending on whether you count me as the current maintainer; I don't). Mark started it, Jim took over it, then Rob. Now I'm doing the majority of the code; though Rob's still the maintainer, handles releases, documentation, and the website (along with Chris), and writes a significant amount of code. I just write code (and an occasional rant).
I can't say how Mark or Jim or Rob got involved with Gaim; I wasn't there for it and don't know the whole stories. The reason I got involved with Gaim was because it sucked. The GNOME panel applet version didn't even compile; the first patch I put in to Gaim was to get it to at least compile. I'm currently in the process of getting the rest of it to do what I want it to do.
Right now there are three or four people who Rob and I would be quite comfortable giving the code to if we wanted to quit. I don't think it's ever really been an issue; there will always be someone there making it better.
I think, possibly unfortunately depending on perspective, that bjb is right: Windows for some reason doesn't attract people who are willing and able to write open source code. Whatever the reason, it ends up that Windows projects are hard to pass on. But it's not solely limited to Windows open source projects. Any project where people aren't submitting patches is hard to pass on.
What I would do, if I didn't know who to pass the code on to, is to announce my retirement, and start keeping a public list of feature requests made by you and other users (SourceForge is great for this, hint hint, plug plug). Someone may eventually put in patches, and you can pass the project off to them. If no one does, you still can write code for it when you have time and interest. Of course, if you can't come up with any feature requests, and neither can anyone else, does it really need to continue to be worked on, other than bug fixes?
The long and thin design also might not be as comfortable as the Gulfstream.
Er. The Gulfstream V is a Gulfstream IV with an extended fuselage. The Gulfstream IV is somewhat narrow, and they haven't widened it because they want to keep the larger windows. (And when I say larger, I mean larger; the windows are ovular and about 2 to 2.5 times the size of the normal ones you see on commercial jets.)
The larger windows are part of an old series of planes with the FAA used to allow but have since grandfathered. You can't use the larger windows on new planes unless if you're using the exact same fuselage as was previously allowed; extending the fuselage was allowed, however, hence the Gulfstream V, which has a larger capacity and I think also more powerful engines, but is just as narrow as a Gulfstream IV. But what a view out of those windows; when you're cruising at 6-12 thousand feet over the coastline of the pacific, it's absolutely gorgeous.
There's other advantages to owning a private jet, in addition to the time advantage, but money isn't one of them. Not having to deal with all the other passengers is a big one. Having your own private movie system, complete with DVD players and VCRs and screens for each seat, is a nice perk. Being able to see flight information, like how high you are, what your ground speed is, and ETA is especially nice. On-board private fax, modem, ethernet, A/C outlets, etc. etc. But like I said, you won't be saving money by having your own jet, no matter how much you travel. Feul, landing fees, storage, maintenance, crew, all add up. Not to mention the millions that you pay just to own the jet. But fortunately, most private jet manufacturers artificially inflate the price over the years, to create inflation. With light to moderate use you'll get most of the cost of the jet back when you sell it, assuming you've paid for regular maintenance, of course.
Most private vehicles tend to be bad investments unless if you're using them to sell rides (e.g. busses, taxis, commercial airlines, trains, etc.). It's much more economical to go commercial than to buy a private jet, no matter how much or how little you use it (similarly, it's much more economical to ride the bus than to own a car). But the time savings and comfort level are phenomenal.
It's possible to change a constant, at least in C. It's a workaround and a bit of a hack, but you can. (Thanks to Peter van der Linden and _Expert_C_Programming_ for this and other wonderful tips.) All you need to do is get a pointer to the constant and change the value pointed at.
const int x = 4;
int *y = &x;
*y = 3;/* x is now 3 */
Yes you get warnings from gcc when you try to compile it (because you're discarding the const qualifier), but it does work. I'm boring, I know.
Magnetic devices prices got driven down as the technology became cheaper and more in-demand.
Sorry, I have to nit-pick. The more in-demand something is the higher the price will be; it's a simple supply/demand curve. Higher demand, higher prices. Higher supply, lower prices. Then eventually the technology gets replaced or otherwise becomes obsolete and demand goes to 0, and the company can't even give the stuff away.
I agree with you on everything else though (though personally I don't think any of the possibilities you listed are likely, especially the remote location idea, but it is still possible).
I think you could reword it and be a little clearer as to what you believe your rights to use the system are
Yeah, well.... I think all the shouting is bound to happen no matter what I post and where, because (surprise surprise) most people don't read the articles before they post comments on Slashdot.
The x86 Linux native client doesn't run on some distros (due to libc incompatibilities and such); both the Linux client and the Java client don't have all the features the Windows clients do, last I checked.
I just answered this in another post in this thread. It's a public server. We can use it however we want. I can telnet to it and send it any kind of garbage I want. If the server happens to understand the garbage and respond to it in a way that's useful to me and other people, that's what the server is there for. I have a right to use it, just like I have a right to connect to Slashdot's web server.
I just thought of something else. AOL puts the servers up publicly on the Internet.
Consider for a moment, an HTTP server that someone has put up publicly on the Internet. Anyone can connect to it and use it however they like; the most useful way is when they use a protocol the server understands. Because it is on the Internet, everyone has a right to use it. Certainly you can firewall out certain people and change the protocol implementation to try and discourage certain clients, but the basic fact is, if you have a public server, anyone can use it.
Why should it be any different for an AIM server? Just because they've spent money developing it and running it doesn't mean that it's any different. They can change the protocol implementation as they want, but we still have a right to use a public server. Isn't that what the Internet is all about?
The only thing that might not be acceptible then is whether or not reverse engineering is legal. I won't get into that.
Read in context. AOL giving us TOC means we have every right to use the TOC servers, until AOL takes them away (which they're not likely to do). It doesn't mean we have a right to use the Oscar servers; but it does mean we have a right to use the service. The point is, using Oscar is not a drain on the resources.
Read it again, closer this time. I'm not saying we have a right to use it; never did I say that it was ours or that AOL couldn't stop us or anything like that. I'm saying that resources are not an issue. AOL does not block clients because it is a drain on the resources; that's all I'm saying.
Maybe AOL isn't worried about trying to block TOC because they realize nobody is bothering to use it?
I can name at least 5 clients off the top of my head that use TOC, not counting Gaim. Kaim, EveryBuddy, TiK, Kit, and jaim. People are trying to use it, because it's what they've been given. It works, marginally. It doesn't have features you'd expect from a "real" AIM client, like the ability to request file transfers or the ability to get users' away messages (both of which are a *really* big deal to some people).
I'm not saying we have a right to the servers; it is a priviledge. What I'm saying is, it's not the resources that are the issue. If the resources were the issue, they wouldn't allow TOC at all.
I can't help thinking "Buddy, get a hair cut and get a real job.":)
Funny, my dad said the same thing just yesterday.:)
You'd be amazed how big a deal getting away messages are. The two most often requested features for Gaim before it was able to use Oscar? Getting away messages and file transfer, in that order. It's a really big deal to a lot of people - and often times, away messages are important.
The whole point of Instant Messeging is presense notification, including when a buddy is available to be talked to (basically whether they're online and not away). People put all kinds of information in their away messages, from song quotes, to what they're doing, to messages that say "I'm not really away" (I have seen an away message that said this, literally). Why have an away message without being away? It's like Occ. or DND in ICQ; to let your friends know you're there, but don't really have time for chit-chat, only important stuff.
As for file transfer.... Yeah, it's kinda pointless, but it's a convenience thing. You've already got AIM open, and you're talking about this MP3, might as well send it over. It ends up being really convenient sometimes, too, especially for people that don't want to set up an ftp/http server, or want to send files to Windows users (who probably don't have scp installed), and who use Yahoo and so can't send/receive that 6.1 MB MP3 (since Yahoo limits mailbox size to 6 MB).
So, once again, while TOC has the basics down, it's missing "important" things:)
If AOL had a client for Linux, *BSD, Solaris, AIX, etc., etc., that could support all of the features the Windows client supports, then I don't think that any of this would really be a big deal; we wouldn't be able to claim as much to it. But in all honesty, the only way they can really provide a client for every OS Gaim (and other open-source AIM clients) supports is by providing source - something I'm sure AOL isn't about to do.
I think AOL is actually benefiting from our (open source AIM clients') existence; we're increasing their userbase. I bet a rather large percentage of our users are people who use Unix but wouldn't use AIM if it weren't for their parents/siblings/spouse/children/friends. So they're not going to get rid of us until they know there's alternatives. And it isn't economically viable to provide alternatives themselves.
So yes. It is a shaky argument, that we should be able to use Oscar because they let us use TOC. But I don't think AOL is going to argue the point when it's just an AIM client. When it's a server, though, that's personal:)
Somehow, I don't think this would work too well, since Gaim and Jabber are both GPL (Jabber is dual-licensed). AOL probably wouldn't be too keen on us redistributing the code that we'd have licensed. If the were to provide it under the GPL, then we'd be more than happy to use it. But if they won't allow that, then we won't be able to use it.
There are a lot of reasons to use AIM over IRC; one of the main reasons, the primary benefit of Instant Messeging in general, is presense notification. How many people on IRC change their nick from person to person_Asleep when they go to bed? Nearly everyone that actually *uses* IRC? How many other tricks to people have on IRC to try and emulate an AIM buddy list?
The point of IRC was never the same as the point of AIM. IRC is a collaboration tool. It's not very useful beyond chat rooms, though it's *very* useful for that. AIM isn't very useful for chat rooms, but it's incredibly useful for knowing when your friends are online, when they're available to talk to you, for actually talking to them in a "private" setting (most people are blissfully unaware that AOL has the right to monitor your conversation), etc.
IRC has its uses, and so does Instant Messaging. Now if only everyone would use Jabber (and Jabber's IRC transport) instead of AIM....
Now, if AOL wished to remove TOC, they would have to do this by disabling capabilites at the server. Is this not what they have already done?
Eh. Not quite. They've removed a lot of functionality, but not the core functionality. Buddy lists and messaging still works. You can still connect and be seen by normal AIM users as being online. But there's a lot more to AIM than just that. That's why TOC isn't something you'd want as a permanent solution. It's not a replacement for Oscar. But at least it works.
Actually I'm kind of curious why AOL doesn't drop TOC completely, take the servers offline. They're not making money off of it. It's only costing them money, having to run and maintain the servers (which they do rather poorly, btw; the TOC servers go down for days at a time on a regular basis). But I'm not going to question it too much; I wouldn't want to give anyone any ideas. (It would *really* suck if in addition to the Oscar blocks TOC wasn't even opeartional - every unofficial AIM client would be dead in the water.)
With no shoes. And back then, books were *books*, not this wimpy paperback crap. My backpack was 50 lbs on a day where I only had 5 hours of homework, and normally I had 10. And I still had time to do all my chores, make dinner for my family (and I have 8 brothers and 6 sisters!), work to help support my family, and learn how to play three different instruments! On top of that, I was captain of the football team, student body president, and valedictorian!
Each client sends a client string specifying which version it is. AOL can then check the records, compare the response, and if it's valid for the given client string, then it'll accept it. All they have to do is have a copy of every binary they've ever made public, which I'm sure they do, at least internally.
It can't find out what version it is just by looking at the file, unfortunately. Also, it needs to use one of two specific versions in order for this to work at all - 3.5 or 4.1. 4.3 will not work because it has server-side buddy lists, and when those are in effect then presence (updates for who's online) doesn't work with the current libfaim. So you have to have either 3.5 or 4.1 - and right now it's hard-coded that you have to have 3.5.
An MD5 server is not feasible. There are two ways to implement said server. One is to distribute the MD5 encoded parts of the binary, and the other is to just distribute parts of the binary and let each client encode it. Each one is not possible.
The first one is not possible because the demand on the server is simply too great. Doing the number of requests required, in the amount of time each request needs to take, isn't possible. Each request needs to be submitted, processed, and completed, before the client's connection to the AIM server times out (which is actually relatively quickly - I don't have an exact value for you but it's in the range of 20 or 30 seconds (rough guess)). For people who aren't on high-bandwidth connections (modem users) this isn't an option.
The second one isn't possible for the same reason, but the amount of data transfered is greater (potentially up to 2^32 bytes (or whatever the size of aim.exe is)), in addition to the fact that it's probably not legal to distribute parts of the binary in this way.
So while an MD5 server for this would be really neat, it's not feasible. Sorry.
Slashdot Mirror
Uh.... AIM servers (and ICQ too now) listen on every port, so you can't block their AIM and ICQ access based on port alone. And actually the AIM client's network setup will bounce through random ports searching for one that works, so they don't even have to try that hard to get around it. I don't know about MSN and Yahoo for certain but I think they only listen on one port. But if you think you're blocking AIM, you're wrong.
I don't feel that's ironic or hypocritical at all.
The worst thing that can happen for Microsoft is for users to have viable alternatives. Compatibility increases the chances of that. Source improves compatibility.
Gaim has actually transferred hands two or three times (depending on whether you count me as the current maintainer; I don't). Mark started it, Jim took over it, then Rob. Now I'm doing the majority of the code; though Rob's still the maintainer, handles releases, documentation, and the website (along with Chris), and writes a significant amount of code. I just write code (and an occasional rant).
I can't say how Mark or Jim or Rob got involved with Gaim; I wasn't there for it and don't know the whole stories. The reason I got involved with Gaim was because it sucked. The GNOME panel applet version didn't even compile; the first patch I put in to Gaim was to get it to at least compile. I'm currently in the process of getting the rest of it to do what I want it to do.
Right now there are three or four people who Rob and I would be quite comfortable giving the code to if we wanted to quit. I don't think it's ever really been an issue; there will always be someone there making it better.
I think, possibly unfortunately depending on perspective, that bjb is right: Windows for some reason doesn't attract people who are willing and able to write open source code. Whatever the reason, it ends up that Windows projects are hard to pass on. But it's not solely limited to Windows open source projects. Any project where people aren't submitting patches is hard to pass on.
What I would do, if I didn't know who to pass the code on to, is to announce my retirement, and start keeping a public list of feature requests made by you and other users (SourceForge is great for this, hint hint, plug plug). Someone may eventually put in patches, and you can pass the project off to them. If no one does, you still can write code for it when you have time and interest. Of course, if you can't come up with any feature requests, and neither can anyone else, does it really need to continue to be worked on, other than bug fixes?
Er. The Gulfstream V is a Gulfstream IV with an extended fuselage. The Gulfstream IV is somewhat narrow, and they haven't widened it because they want to keep the larger windows. (And when I say larger, I mean larger; the windows are ovular and about 2 to 2.5 times the size of the normal ones you see on commercial jets.)
The larger windows are part of an old series of planes with the FAA used to allow but have since grandfathered. You can't use the larger windows on new planes unless if you're using the exact same fuselage as was previously allowed; extending the fuselage was allowed, however, hence the Gulfstream V, which has a larger capacity and I think also more powerful engines, but is just as narrow as a Gulfstream IV. But what a view out of those windows; when you're cruising at 6-12 thousand feet over the coastline of the pacific, it's absolutely gorgeous.
There's other advantages to owning a private jet, in addition to the time advantage, but money isn't one of them. Not having to deal with all the other passengers is a big one. Having your own private movie system, complete with DVD players and VCRs and screens for each seat, is a nice perk. Being able to see flight information, like how high you are, what your ground speed is, and ETA is especially nice. On-board private fax, modem, ethernet, A/C outlets, etc. etc. But like I said, you won't be saving money by having your own jet, no matter how much you travel. Feul, landing fees, storage, maintenance, crew, all add up. Not to mention the millions that you pay just to own the jet. But fortunately, most private jet manufacturers artificially inflate the price over the years, to create inflation. With light to moderate use you'll get most of the cost of the jet back when you sell it, assuming you've paid for regular maintenance, of course.
Most private vehicles tend to be bad investments unless if you're using them to sell rides (e.g. busses, taxis, commercial airlines, trains, etc.). It's much more economical to go commercial than to buy a private jet, no matter how much or how little you use it (similarly, it's much more economical to ride the bus than to own a car). But the time savings and comfort level are phenomenal.
"Cars are only dangerous if they can move."
It's possible to change a constant, at least in C. It's a workaround and a bit of a hack, but you can. (Thanks to Peter van der Linden and _Expert_C_Programming_ for this and other wonderful tips.) All you need to do is get a pointer to the constant and change the value pointed at.
/* x is now 3 */
const int x = 4;
int *y = &x;
*y = 3;
Yes you get warnings from gcc when you try to compile it (because you're discarding the const qualifier), but it does work. I'm boring, I know.
Sorry, I have to nit-pick. The more in-demand something is the higher the price will be; it's a simple supply/demand curve. Higher demand, higher prices. Higher supply, lower prices. Then eventually the technology gets replaced or otherwise becomes obsolete and demand goes to 0, and the company can't even give the stuff away.
I agree with you on everything else though (though personally I don't think any of the possibilities you listed are likely, especially the remote location idea, but it is still possible).
-----
include/asm-mips/mmu_context.h:/* Fuck. The f-word is here so you can grep for it :-) */
-----
No, but it should be :) Enter Jabber.
I think you could reword it and be a little clearer as to what you believe your rights to use the system are
Yeah, well.... I think all the shouting is bound to happen no matter what I post and where, because (surprise surprise) most people don't read the articles before they post comments on Slashdot.
-----
The x86 Linux native client doesn't run on some distros (due to libc incompatibilities and such); both the Linux client and the Java client don't have all the features the Windows clients do, last I checked.
-----
I just answered this in another post in this thread. It's a public server. We can use it however we want. I can telnet to it and send it any kind of garbage I want. If the server happens to understand the garbage and respond to it in a way that's useful to me and other people, that's what the server is there for. I have a right to use it, just like I have a right to connect to Slashdot's web server.
-----
Consider for a moment, an HTTP server that someone has put up publicly on the Internet. Anyone can connect to it and use it however they like; the most useful way is when they use a protocol the server understands. Because it is on the Internet, everyone has a right to use it. Certainly you can firewall out certain people and change the protocol implementation to try and discourage certain clients, but the basic fact is, if you have a public server, anyone can use it.
Why should it be any different for an AIM server? Just because they've spent money developing it and running it doesn't mean that it's any different. They can change the protocol implementation as they want, but we still have a right to use a public server. Isn't that what the Internet is all about?
The only thing that might not be acceptible then is whether or not reverse engineering is legal. I won't get into that.
-----
Read in context. AOL giving us TOC means we have every right to use the TOC servers, until AOL takes them away (which they're not likely to do). It doesn't mean we have a right to use the Oscar servers; but it does mean we have a right to use the service. The point is, using Oscar is not a drain on the resources.
-----
Read it again, closer this time. I'm not saying we have a right to use it; never did I say that it was ours or that AOL couldn't stop us or anything like that. I'm saying that resources are not an issue. AOL does not block clients because it is a drain on the resources; that's all I'm saying.
-----
I can name at least 5 clients off the top of my head that use TOC, not counting Gaim. Kaim, EveryBuddy, TiK, Kit, and jaim. People are trying to use it, because it's what they've been given. It works, marginally. It doesn't have features you'd expect from a "real" AIM client, like the ability to request file transfers or the ability to get users' away messages (both of which are a *really* big deal to some people).
I'm not saying we have a right to the servers; it is a priviledge. What I'm saying is, it's not the resources that are the issue. If the resources were the issue, they wouldn't allow TOC at all.
I can't help thinking "Buddy, get a hair cut and get a real job." :)
Funny, my dad said the same thing just yesterday. :)
-----
The whole point of Instant Messeging is presense notification, including when a buddy is available to be talked to (basically whether they're online and not away). People put all kinds of information in their away messages, from song quotes, to what they're doing, to messages that say "I'm not really away" (I have seen an away message that said this, literally). Why have an away message without being away? It's like Occ. or DND in ICQ; to let your friends know you're there, but don't really have time for chit-chat, only important stuff.
As for file transfer.... Yeah, it's kinda pointless, but it's a convenience thing. You've already got AIM open, and you're talking about this MP3, might as well send it over. It ends up being really convenient sometimes, too, especially for people that don't want to set up an ftp/http server, or want to send files to Windows users (who probably don't have scp installed), and who use Yahoo and so can't send/receive that 6.1 MB MP3 (since Yahoo limits mailbox size to 6 MB).
So, once again, while TOC has the basics down, it's missing "important" things :)
-----
If AOL had a client for Linux, *BSD, Solaris, AIX, etc., etc., that could support all of the features the Windows client supports, then I don't think that any of this would really be a big deal; we wouldn't be able to claim as much to it. But in all honesty, the only way they can really provide a client for every OS Gaim (and other open-source AIM clients) supports is by providing source - something I'm sure AOL isn't about to do.
I think AOL is actually benefiting from our (open source AIM clients') existence; we're increasing their userbase. I bet a rather large percentage of our users are people who use Unix but wouldn't use AIM if it weren't for their parents/siblings/spouse/children/friends. So they're not going to get rid of us until they know there's alternatives. And it isn't economically viable to provide alternatives themselves.
So yes. It is a shaky argument, that we should be able to use Oscar because they let us use TOC. But I don't think AOL is going to argue the point when it's just an AIM client. When it's a server, though, that's personal :)
-----
Somehow, I don't think this would work too well, since Gaim and Jabber are both GPL (Jabber is dual-licensed). AOL probably wouldn't be too keen on us redistributing the code that we'd have licensed. If the were to provide it under the GPL, then we'd be more than happy to use it. But if they won't allow that, then we won't be able to use it.
-----
The point of IRC was never the same as the point of AIM. IRC is a collaboration tool. It's not very useful beyond chat rooms, though it's *very* useful for that. AIM isn't very useful for chat rooms, but it's incredibly useful for knowing when your friends are online, when they're available to talk to you, for actually talking to them in a "private" setting (most people are blissfully unaware that AOL has the right to monitor your conversation), etc.
IRC has its uses, and so does Instant Messaging. Now if only everyone would use Jabber (and Jabber's IRC transport) instead of AIM....
-----
Eh. Not quite. They've removed a lot of functionality, but not the core functionality. Buddy lists and messaging still works. You can still connect and be seen by normal AIM users as being online. But there's a lot more to AIM than just that. That's why TOC isn't something you'd want as a permanent solution. It's not a replacement for Oscar. But at least it works.
Actually I'm kind of curious why AOL doesn't drop TOC completely, take the servers offline. They're not making money off of it. It's only costing them money, having to run and maintain the servers (which they do rather poorly, btw; the TOC servers go down for days at a time on a regular basis). But I'm not going to question it too much; I wouldn't want to give anyone any ideas. (It would *really* suck if in addition to the Oscar blocks TOC wasn't even opeartional - every unofficial AIM client would be dead in the water.)
-----
With no shoes. And back then, books were *books*, not this wimpy paperback crap. My backpack was 50 lbs on a day where I only had 5 hours of homework, and normally I had 10. And I still had time to do all my chores, make dinner for my family (and I have 8 brothers and 6 sisters!), work to help support my family, and learn how to play three different instruments! On top of that, I was captain of the football team, student body president, and valedictorian!
-----
Each client sends a client string specifying which version it is. AOL can then check the records, compare the response, and if it's valid for the given client string, then it'll accept it. All they have to do is have a copy of every binary they've ever made public, which I'm sure they do, at least internally.
It can't find out what version it is just by looking at the file, unfortunately. Also, it needs to use one of two specific versions in order for this to work at all - 3.5 or 4.1. 4.3 will not work because it has server-side buddy lists, and when those are in effect then presence (updates for who's online) doesn't work with the current libfaim. So you have to have either 3.5 or 4.1 - and right now it's hard-coded that you have to have 3.5.
An MD5 server is not feasible. There are two ways to implement said server. One is to distribute the MD5 encoded parts of the binary, and the other is to just distribute parts of the binary and let each client encode it. Each one is not possible.
The first one is not possible because the demand on the server is simply too great. Doing the number of requests required, in the amount of time each request needs to take, isn't possible. Each request needs to be submitted, processed, and completed, before the client's connection to the AIM server times out (which is actually relatively quickly - I don't have an exact value for you but it's in the range of 20 or 30 seconds (rough guess)). For people who aren't on high-bandwidth connections (modem users) this isn't an option.
The second one isn't possible for the same reason, but the amount of data transfered is greater (potentially up to 2^32 bytes (or whatever the size of aim.exe is)), in addition to the fact that it's probably not legal to distribute parts of the binary in this way.
So while an MD5 server for this would be really neat, it's not feasible. Sorry.