OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   Re: [xml-dev] Typing and paranoia

[ Lists Home | Date Index | Thread Index ]

Alaric B. Snell wrote:

>>REALLY want to be able to deal with the bits on the wire and REALLY like
>>the open-ness and interoperability that gives us.  Others really REALLY
>>want to take the bits on the wire away and present us instead with an
>>API that has 117 entry points averaging 5 arguments and try to convince
>>us that this is somehow equivalent. 
> I don't think anyone wants that, silly!
> What people seem to want is to be able to have information in program A and 
> to send it over thet network to program B. This can be done perfectly well 
> without needing to do much more than say "Runtime library, send this over 
> there!" - the main issue is compatability of object models rather than the 
> precise bit streams 

This is precisely where we disagree.  For years vendors have been 
telling me "don't bother your pretty little head over the bits on the 
wire, just put the data in here and it will come out there."  Except 
when it doesn't.  Except when I have the audacity to use a programming 
language or operating system that doesn't have the library.  Except when 
I don't have the budget to acquire the library.  If there's a library 
there I'll cheerfully use it, but I want guaranteed access to my own 
data in an open format as a basic condition of being willing to play.  I 
think I'm pretty well in line with the CIOs of the Global 2000 in my 
feelings on this; these guys are all covered in scars from the API wars.

> This is what ASN.1 provides, and XDR to a lesser extent, and CORBA to a 
> similar extent with a bit more overhead, and RMI provides with certain 
> caveats (THAT GET ON MY NERVES) within the Java environment...

You're proving my point.  If all these things exist and they solve the 
problem, why is everyone so cheery about the idea of pumping XML 
messages back and forth, especially given that every one of the above 
has a wire format that is probably way more efficient.

> In case you hadn't noticed, there was a lot of interoperability before XML 
> came around:
> 1) The international telephone network
> 2) The Internet (from IP up to the WWW)
> 3) Image file formats
> 4) CSV and TSV
> 5) Zip files
> 6) Filesystems

Game, set, and match to me.  All of these - every one - are based on 
open, documented data formats.  It's the ***>ONLY WAY<*** to get 
interoperability.  Show me the bits on the wire, please.

> Throw off those rose tinted glasses! We were interoperating quite happily 
> before XML came along, thanks. Maybe you spent most of those 21 years in 
> Microsoft environments which don't interoperate well with the rest, but the 
> rest of us were happily moving code and data between (in my case) various 
> flavours of unix, DOS, RISC OS, and Windows.

Gasp choke... must have hallucinated all those weeks and months I spent 
porting Unix applications from SunOS to Solaris to AIX to RISCOS to 
Xenix, and the *years* of my life I spent fighting through the 16->32 
bit transition (char far * anyone?).  Unix APIs are interoperable to the 
exact degree that they have a common codebase, otherwise not.  Counting 
on a common codebase is a lousy way to design the future.

>>But you're not going to take the bits on the wire away from us without a
>>huge messy noisy fight down to the last ditch. -Tim
> Hahahah.

So far, your arguments are losing - not on xml-dev, but in the 
marketplace, where it matters. -Tim


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS