I took another look at this new build and posted some commens on my blog (to much to fit in this little comment box: http://robert.accettura.com/archives/2005/05/25/aim-triton-take-2/
Seeing as how the previous version didn't even let me sign on, this is quite an improvement. However, I'm just wondering a couple of things.
First and foremost, what happened to control character support? For example, users were able to put '%n' into links, thus enabling sites like <A HREF='http://www.subprofile.com/'>Subprofile.com</A> and <A HREF='http://www.slapm.com/'>SLAPM.com</A> to enhance the end-user experience with things like quizzes and random quotes (with Subprofile) or enhanced privacy control (with SLAPM). This is especially disconcerting for me as a user of SLAPM.
Next, why is it that there is no way of closing the tray icon? Hijacking desktop space without an easy way to exit is very rude. I for one would not care for a tray icon at all, let alone one that is persistent even when I am not using AIM. Maybe make the tray icon an option?
And finally, the window behavior is a little bit weird. Whenever I start Triton, it starts up all the way to the right side of my screen, and I have to right click its taskbar icon and select 'Move' to bring it into view. That, and the message window seems extremely cluttered. There should definitely be a way to enable classic view, or something like that. The way it is now, Triton definitely is going to turn people away from AIM. I sincerely hope that you make it better for the sake of all of your users.
'bout the AIM6 code : you said it would be "portable".
As a programmer, let me ask some questions and give some tips, though i take it you have much more skill than me.
Questions : - in which language is the AIM6 code written? - do you use MFC very? or whatever? - is the code separated in, let's say, FileTransferLibrary.(a, so or dll, whatever), MessageLibrary.(whatever), and Interface.exe? Or whatever, i think you catch the drift.
Tips/wishlist :P : - Did you think about writing AIM in a .net language? Or better, with Mono? That way, you can at least ensure that the code compiles on Mono AND Official .net, and have only to recompile to get the client - yes, i know, dumb idea. - Multiplatform gui toolkit : wxWidgets - www.wxwidgets.org - Soon-to-be multiplatform toolkit, though I doubt they'll get anything fast using winelib : http://smartwin.sourceforge.net/ - Why in the nine layers of hell don't you do all versions with a single toolkit? Surely you can think of something. Hell, if the code is very modular, and i underline VERY, you could have the interface code in one module only, and only change THAT module from time to time if needed.
Hey, I really like how the Triton beta is looking, much less bloat than 5.9 (which I ditched in favor of GAIM).
However, some very crucial features for me are still missing, AIM auto-logging and of tabbed messaging. Also, I think something like GAIM's auto away message on computer Idle would be really useful, just some input for you!
AIM Triton is written in C++ and Javascript. We don't use MFC. Similar to what you say, the code is divided into user interface, business logic, and protocol libraries; we do use a single UI toolkit (our own Boxely tookit). We don't use .NET (perf in .NET is still not very good) although we support .NET via COM interop.
I'm also a programmer and I'd like to thank you for building a "pure" app; downloading a huge framework just to make a small app work is more annoying then anything - IMHO; Java and .Net is good on servers and big "enterprise" software.
Now I don't know if you plan to make a Un*x version but I'll say this anyway: I don't think you should - unless you want to have fun - don't even release a library. A nice document that describes the protocol would be very nice and there's a simple reason to it: (Background info, I'm a Linux user) GAIM, Kopete and other IM client in the Unix world already has their own logic layout- if you released a library they would have to adapt their code - and I don't think they want to rip hours of hard work... But if they have the proto documentation they can continue to implant it the way they started and that would allow them to build a version faster. Again this is my own 2cents; I could be very wrong.. or too tired to see things straight.
Hope you fix coplands memory leak. it's sending my aolserivcehost.exe up towards 100% also why aren't you putting a save screen name and password into the sign on? i asked the beta staff and they told me no plans are to add it in? its a pain in the butt to keep typing it in.
Replying to nomailnochat: Saved password is something we are aware is missing, and are working on it to get it into Triton. In terms of memory usage, Justin said it best when beta 1 was released, "Note that this is a very early beta, of a product that is a complete rewrite... it's not very polished in a lot of places. The standard disclaimers about beta software definitely apply here." We are constantly making tweaks and adjustments to perf and memory, and hopefully with each successive release things will improve.
One additional note/opinion regarding .Net. Perf with .Net is not my only objection, but debugging really is a pain with no edit/continue and very painful experience attaching to process to debug across languages. In my personal experience C# is a terrific language for quick development, but it is not ready for prime time when writing a super robust app.
16 comments:
Sounds nice, but when will Triton be available, if ever, for lower Windows versions? (ie. 2000 professional)
I took another look at this new build and posted some commens on my blog (to much to fit in this little comment box:
http://robert.accettura.com/archives/2005/05/25/aim-triton-take-2/
Seeing as how the previous version didn't even let me sign on, this is quite an improvement. However, I'm just wondering a couple of things.
First and foremost, what happened to control character support? For example, users were able to put '%n' into links, thus enabling sites like <A HREF='http://www.subprofile.com/'>Subprofile.com</A> and <A HREF='http://www.slapm.com/'>SLAPM.com</A> to enhance the end-user experience with things like quizzes and random quotes (with Subprofile) or enhanced privacy control (with SLAPM). This is especially disconcerting for me as a user of SLAPM.
Next, why is it that there is no way of closing the tray icon? Hijacking desktop space without an easy way to exit is very rude. I for one would not care for a tray icon at all, let alone one that is persistent even when I am not using AIM. Maybe make the tray icon an option?
And finally, the window behavior is a little bit weird. Whenever I start Triton, it starts up all the way to the right side of my screen, and I have to right click its taskbar icon and select 'Move' to bring it into view. That, and the message window seems extremely cluttered. There should definitely be a way to enable classic view, or something like that. The way it is now, Triton definitely is going to turn people away from AIM. I sincerely hope that you make it better for the sake of all of your users.
-Pete
I apologize for the stray HTML in the previous comment. The links should have read:
http://www.subprofile.com/ - Subprofile
http://www.slapm.com/ - SLAPM
Thanks!
'bout the AIM6 code : you said it would be "portable".
As a programmer, let me ask some questions and give some tips, though i take it you have much more skill than me.
Questions :
- in which language is the AIM6 code written?
- do you use MFC very? or whatever?
- is the code separated in, let's say, FileTransferLibrary.(a, so or dll, whatever), MessageLibrary.(whatever), and Interface.exe? Or whatever, i think you catch the drift.
Tips/wishlist :P :
- Did you think about writing AIM in a .net language? Or better, with Mono? That way, you can at least ensure that the code compiles on Mono AND Official .net, and have only to recompile to get the client - yes, i know, dumb idea.
- Multiplatform gui toolkit : wxWidgets - www.wxwidgets.org
- Soon-to-be multiplatform toolkit, though I doubt they'll get anything fast using winelib : http://smartwin.sourceforge.net/
- Why in the nine layers of hell don't you do all versions with a single toolkit? Surely you can think of something. Hell, if the code is very modular, and i underline VERY, you could have the interface code in one module only, and only change THAT module from time to time if needed.
Hey, I really like how the Triton beta is looking, much less bloat than 5.9 (which I ditched in favor of GAIM).
However, some very crucial features for me are still missing, AIM auto-logging and of tabbed messaging. Also, I think something like GAIM's auto away message on computer Idle would be really useful, just some input for you!
Forget the tabbed messaging comment, I am dumb and didn't realize it was there.
Replying to "jsrlepage":
AIM Triton is written in C++ and Javascript. We don't use MFC. Similar to what you say, the code is divided into user interface, business logic, and protocol libraries; we do use a single UI toolkit (our own Boxely tookit). We don't use .NET (perf in .NET is still not very good) although we support .NET via COM interop.
I'm also a programmer and I'd like to thank you for building a "pure" app; downloading a huge framework just to make a small app work is more annoying then anything - IMHO; Java and .Net is good on servers and big "enterprise" software.
Now I don't know if you plan to make a Un*x version but I'll say this anyway: I don't think you should - unless you want to have fun - don't even release a library. A nice document that describes the protocol would be very nice and there's a simple reason to it: (Background info, I'm a Linux user) GAIM, Kopete and other IM client in the Unix world already has their own logic layout- if you released a library they would have to adapt their code - and I don't think they want to rip hours of hard work... But if they have the proto documentation they can continue to implant it the way they started and that would allow them to build a version faster. Again this is my own 2cents; I could be very wrong.. or too tired to see things straight.
Hope you fix coplands memory leak. it's sending my aolserivcehost.exe up towards 100% also why aren't you putting a save screen name and password into the sign on? i asked the beta staff and they told me no plans are to add it in? its a pain in the butt to keep typing it in.
replying to lyoshil :
in one way, you are right, kopete, gaim et al. already have their interface logic.
you are however forgetting some things :
1- what about AIM having a "universal" interface throughout the OSs?
2- GAIM very sucks at filetransfering things.
but you do have a good point.
Replying to nomailnochat:
Saved password is something we are aware is missing, and are working on it to get it into Triton. In terms of memory usage, Justin said it best when beta 1 was released, "Note that this is a very early beta, of a product that is a complete rewrite... it's not very polished in a lot of places. The standard disclaimers about beta software definitely apply here." We are constantly making tweaks and adjustments to perf and memory, and hopefully with each successive release things will improve.
One additional note/opinion regarding .Net. Perf with .Net is not my only objection, but debugging really is a pain with no edit/continue and very painful experience attaching to process to debug across languages. In my personal experience C# is a terrific language for quick development, but it is not ready for prime time when writing a super robust app.
What is the target date for triton? end of the year? next spring? or what
I may have missed it, but will the final version have Logging, GAIM/DeadAIM-esque?
what the heck y cant you change you profile on triton!!!!!!!!!urghhhhhhhhhhhhh
grr!! why cant i have an icon?! from iconator.com?!!?? grrr wht happend to my profile
Post a Comment