|
|
OpenH323 Project Kickoff Letter | ||
|
Mailing list Mirrors |
The following letter was sent on December 5 1998 to a small number of people who had contacted me regarding working on the OpenH323 project. It was intended to kickoff the OpenH323 project by explaining where the project had come from, and raising some issues to get everyone thinking along the same lines. If you want join the fun, I suggest you read the letter, join the mailing list and start thinking! To those interested in participating in an effort to create Open Source H.323 Protocol suite. 5 December 1998 Dear Internet User, You've received this email because you have previously expressed interest in participating in an Open Source effort to create a H.323 protocol stack call the Linux Voxilla H.323 Project. I thought you might be interested in hearing about some recent changes to this project, and being involved in getting the project off the ground. History I started this project after seeing an article on http://slashdot.org about Rich Bodo's Voxilla project at http://www.voxilla.org. My company, Equivalence, had been contemplating starting an Open Source H.323 project for some time, and this looked like a good way to start it off. Rich was very supportive of the whole idea, and I created a new web site at http://www.voxillaequival.net to be the home site for the Linux Voxilla H.323 Project. Unfortunately, things seem to stall soon after that, and not a lot of progress has been made since then. Also unfortunately, Rich Bodo seems to have dropped out of contact. The moderated "voxilla" mailing list does not seem to be be issuing any new messages, and Rich isn't answering his emails. http://www.voxilla.org hasn't been changed for a while either. (UPDATE - 7 Dec 98: Rich received a copy of this message, and is now back in contact. Click here for more information) So, I've decided that the time has come to move on. We wish Rich well, and I would like to personally thank him for making me get off my butt and do something about Open Source H.323. If you ever read this email, Rich, please don't take this as a criticism of your efforts or ideas, or think we are trying to throw you out of the project. You will always be welcome! OpenH323 The Linux Voxilla H.323 Project has been renamed OpenH323 to reflect the current trend in Open Source projects, and has a new home at http://www.openh323.equival.net. This site has been updated to reflect the new name and the new contact email addresses. The old http://www.voxilla.equival.net URL still works, but just maps straight to the new site. The domain name "openh323.org" has also been applied for by Equivalence on behalf of the OpenH323 project and http://openh323.org will be up and running shortly. A new, unmoderated, majordomo mailing list has also been created for this project. Thist list is called "openh323" and is hosted at majordom@equival.net You have already been subscribed to this list - to get more information see http://www.openh323.equival.net/list.html. Getting started I would like to discuss some issues regarding the start of OpenH323 with some people who have time to spare and are interested in contributing to the early, hard stages of getting something going. The original concept behind Equivalence being involved in an OpenH323 project was that we would make our multi-platform, C++ code library Open Source, and hence freely available to all OpenH323 contributors. We have developed this code library, called PwLib, over the past four years, and it allows developers to write applications that are portable across Linux x86, Linux PPC, Solaris Sparc, Windows 95/98, Windows NT, and FreeBSD x86. By "portable", we mean just recompile the application source code, link with the appropriate platform specific library, and run the program. No ifdefs, no other changes required. It fully supports all major system facilities such as multithreading (it does cooroperative multi-tasking on platforms that don't have threads) and TCP/IP sockets, and has a full C++ container library with support for abstract data structures such as queues, arrays and sorted lists. We have found this library invaluable, and we use it in all of our commercial products. It compiles with MSVC C++ 5.0 on Windows, and uses GCC with GLIBC on the various Unices. Included with this library would be our ASN parser which is capable of converting the standard H.323 and H.225 ASN specifications in C++ classes that can be compiled and linked with the PWLib class library. We have PER encoding and decoding functions, and the resultant code has been used in our own H.323 product, PhonePatch, so we know the H.323 and H.225 PDU encoding/decoding works correctly. Questions for Discussion 1) Is using PwLib a good idea? We think it is, because it gives us platform independence for stuff like threads and because we like C++. It also means we have a "ready to go" ASN compiler (a non-trivial task for those who've tried to write one!). Others might not like it because it is yet more code to come to grips with, and they don't like C++. Also, someone else might have a better ASN compiler that can be thrown into the ring. I'd like to see some discussion of this before we go much further. 2) What sort of approach should we use? I don't think there is any doubt about creating a library, rather than just creating a full-blown app, but should this library be multi-threaded? Single-threaded? What sort of API should be used? People with experience in other H.323 stacks will undoubtedly have some ideas here - let's hear them and get them down on paper (HTML?) where other can see them. Regards Craig Southeren |
||
|