- From: Adi Oltean <aoltean@Exchange.Microsoft.com>
- Date: Thu, 18 May 2000 14:27:28 -0700
- To: "Justin Chapweske" <justin@cyrus.net>
- Cc: <xml-dist-app@w3.org>
- Message-ID: <19398D273324D3118A2B0008C7E9A56909C91FE9@SIT.platinum.corp.microsoft.com>
I agree with your point - and this reminds me a similar problem in the past: when CORBA 1.0 came out it didn't present a semantic standard for marshalling object references, for example. This was just one of the causes for incompatibilies between various ORB vendors - Inter-ORB bridging was hard to realize since these protocols differed in subtle different ways in marshalling objects by reference or by value. But mainly each ORB vendor had its own proprietary protocol. When IIOP came out (four years later, if I remember correctly) it was too late: the existing native intra-ORB protocols at that time were much more efficient and richer in functionality than IIOP (for example IIOP didn't had GC) and this discouraged a wide adoption of IIOP as the unique protocol for comunnication under inter/intra ORB environments. Personally, I think SOAP will face the same dangers in future (unless a couple of smart guys will come early with a good open standard extension of SOAP concerning these issues) Thanks, Adi -----Original Message----- From: Justin Chapweske [mailto:justin@cyrus.net] Sent: Thursday, May 18, 2000 12:26 PM To: Adi Oltean Cc: xml-dist-app@w3.org Subject: Re: XML protocol security Excellent points adi, but I wonder if we can't have both. I seriously doubt that anyone is going to want to sacrifice SOAP's simplicity to add object references and a strong security model, but there needs to be a realization from all of the SOAP advocates that this is a designed limitation. The reason I say this is because if SOAP becomes as popular as many think it will, and it has weaknesses, then the SOAP enthusaists need to be able to swallow their pride and recommend stronger solutions. One of these stronger solutions may very well be a SOAP extension to add capabilities and object references, which leads to my question: Does SOAP's implementation simplicity hold as much value once SOAP has been widely deployed and robust tools have been developed for it? The simplicity is a very strong feature for the early adoption of this technology, but as it becomes more mature are we going to be willing to trade off simplicity for stronger security guarentees? If the feeling is that we would be likely to make some trade-offs, then people should consider very carefully the migration path that will need to be taken from the current SOAP to SOAP-FAT in the future... Hope this gets people to throw some food at each other.... -- Justin Chapweske - Noodler, Cyrus Intersoft http://www.cyrusintersoft.com/
Received on Thursday, 18 May 2000 17:58:44 UTC