Comments on: Why text-based protocols hurt your design http://blog.radvision.com/codeofcontact/2008/02/11/why-text-based-protocols-hurt-your-design/ The ins and outs of V²oIP & IMS programming Thu, 20 Nov 2008 00:06:25 +0000 http://wordpress.org/?v=2.6.3 By: Ran Arad http://blog.radvision.com/codeofcontact/2008/02/11/why-text-based-protocols-hurt-your-design/#comment-6 Ran Arad Tue, 12 Feb 2008 10:35:07 +0000 http://blog.radvision.com/codeofcontact/2008/02/11/why-text-based-protocols-hurt-your-design/#comment-6 My suggestion is twofold: One option is to force text encoding to be rigid. For instance, if IMS would have enforced some rigid encoding standard the same way it enforced Sig-Comp, There might have been a movement towards rigid code. Alternatively, a rigid encoding could replace Sig -Comp by some conversion to binary form. Since we do not live in a perfect world, my other suggestion is that programmers be aware of the effect the parsers tend to have on their design, and take measures to prevent it, like forcing the parser to be unaware of context, or at least resisting the temptation to add logic to the parser. My suggestion is twofold:
One option is to force text encoding to be rigid. For instance, if IMS would have enforced some rigid encoding standard the same way it enforced Sig-Comp, There might have been a movement towards rigid code. Alternatively, a rigid encoding could replace Sig -Comp by some conversion to binary form.
Since we do not live in a perfect world, my other suggestion is that programmers be aware of the effect the parsers tend to have on their design, and take measures to prevent it, like forcing the parser to be unaware of context, or at least resisting the temptation to add logic to the parser.

]]> By: admin http://blog.radvision.com/codeofcontact/2008/02/11/why-text-based-protocols-hurt-your-design/#comment-5 admin Tue, 12 Feb 2008 05:44:29 +0000 http://blog.radvision.com/codeofcontact/2008/02/11/why-text-based-protocols-hurt-your-design/#comment-5 Hi Yan, We've just begun posting here on these blogs. I am sure that in the near future you'll see additional blogs pop up here on our site with varied content from different parts of RADVISION. Hi Yan,

We’ve just begun posting here on these blogs. I am sure that in the near future you’ll see additional blogs pop up here on our site with varied content from different parts of RADVISION.

]]> By: Yan Simkin http://blog.radvision.com/codeofcontact/2008/02/11/why-text-based-protocols-hurt-your-design/#comment-4 Yan Simkin Tue, 12 Feb 2008 04:58:56 +0000 http://blog.radvision.com/codeofcontact/2008/02/11/why-text-based-protocols-hurt-your-design/#comment-4 Continuation: ... and the things started moving. Maybe, this is what's going to happen with SIP and the others? BTW, why are there only TBU guys posting here? I would expect the NBU to contribute a bit as well - they have their wise men :) Thanks, WBR, Yan Continuation:
… and the things started moving. Maybe, this is what’s going to happen with SIP and the others?

BTW, why are there only TBU guys posting here? I would expect the NBU to contribute a bit as well - they have their wise men :)

Thanks,
WBR,
Yan

]]> By: Yan Simkin http://blog.radvision.com/codeofcontact/2008/02/11/why-text-based-protocols-hurt-your-design/#comment-3 Yan Simkin Tue, 12 Feb 2008 04:56:19 +0000 http://blog.radvision.com/codeofcontact/2008/02/11/why-text-based-protocols-hurt-your-design/#comment-3 :-S Hi, Ran! As a one who used to work a lot with different vendors of equipment (SIP and H.323) during the "RVSN" part of my life as an Interop Test Eng, I know exactly what you mean when you're talking about the difference in the implementations of the protocols based on text. But what would you suggest? Eliminating text-based protocols? Just for example, in H.323 world, there was a moment when a first standard for Dual Video has arrived (H.329). Despite being a standard, every vendor implemented this in different way and this was quite a headache for a very long period of time. But, at the end, the things got better :-S Hi, Ran!
As a one who used to work a lot with different vendors of equipment (SIP and H.323) during the “RVSN” part of my life as an Interop Test Eng, I know exactly what you mean when you’re talking about the difference in the implementations of the protocols based on text. But what would you suggest? Eliminating text-based protocols?
Just for example, in H.323 world, there was a moment when a first standard for Dual Video has arrived (H.329). Despite being a standard, every vendor implemented this in different way and this was quite a headache for a very long period of time. But, at the end, the things got better

]]>