Re: Why XML to comunicate Distributed Components?

Dmitry A. Kazakov wrote:
On Wed, 27 Feb 2008 09:17:44 -0800 (PST), carofe wrote:

Well, the purposing of this post is to find good reason about xml
technology as message format.
My question is: Why the communication between components should be
XML?

Why should it be? But you put that question too.

XML has some features as self-described, human-readable, etc...

XML is unreadable, at least by normal human beings. Personally, I am unable
to read it without tools. (A need in tools makes the case.)

, and
there are many tools for validation and parsing.

It is a quite poor design to require communicating parties to parse
anything. The protocol has to be simple enough to recognize by a small
state machine.

Validation is not a proper argument here, because it is not the semantics
which is validated, but the XML itself. On this level there is nothing to
validate if a normal (simple, carefully designed) communication protocol is
used. Semantic validation of the content is an independent issue.

Those features are very useful for Configuration files and
Descriptions file which could be edited by humans.

Of course "file" is itself a "protocol" to communicate to a "file system
component". Clearly, it is a bad idea to communicate this way with
something else...

But, Are those features necessaries to communicate two components? Is
it necessary that communication between two component being human-
readable?

Obviously it has to be readable by parties only. If humans are not among
them, there is no reason to care.

About self-described:
Self-described messages is useful for computing, I mean, validations,
infering, etc... But, Is it necessary the message in Plain Text
format?

I doubt that it is self-described. The only semantic an XML document has is
a hierarchical structure and named components. This is too thin. Well, a
human may deduce some information from that. Another question is whether
this information is correct and shared by other human readers. But as for
the real parties, they would not be able to do that. There has to be some
naming conventions shared and implemented in them in order to support such
descriptions. That looks way too complex and the question is what for?

Which are the benefit of use XML to communicate components?

None.

The real power in XML comes from the X.

If you need to make additions to your protocol, older clients should be able to handle the new fields (by simply ignoring them). Try that with a binary protocol. It *is* possible, but it can become quite troublesome.

--
Daniel Pitts' Tech Blog: <http://virtualinfinity.net/wordpress/>
.

Relevant Pages

  • Re: Why XML to comunicate Distributed Components?
    ... technology as message format. ... Why the communication between components should be ... XML is unreadable, at least by normal human beings. ... Validation is not a proper argument here, because it is not the semantics ...
    (comp.object)
  • Re: Why XML to comunicate Distributed Components?
    ... Why the communication between components should be ... there are many tools for validation and parsing. ... Are those features necessaries to communicate two components? ... Which are the benefit of use XML to communicate components? ...
    (comp.object)
  • Re: Automated WSF argument validation: demo script.
    ... The "right" kind of tool for this job is a COM-accessible XML stream reader ... The script is a demo of automating argument validation ... >> end sub ...
    (microsoft.public.windows.server.scripting)
  • Re: Automated WSF argument validation: demo script.
    ... The "right" kind of tool for this job is a COM-accessible XML stream reader ... The script is a demo of automating argument validation ... >> end sub ...
    (microsoft.public.scripting.wsh)
  • Re: 2 Questions about validations
    ... > In SGML and XML, validity means that a document conforms to the Document ... Validation is a process ...
    (comp.infosystems.www.authoring.html)