RE: [xml-dev] Packaging (was Re: [xml-dev] Interoperability)

["Henrik Frystyk Nielsen" <henrikn@microsoft.com>]

James, thanks for pointing at DIME - you are right on with what it is
intended for.

We just submitted DIME as an Internet Draft and it is now available at
[1]. We also have a mailing list for discussion of DIME - you can find
instructions at [2]. There are a few changes from the spec on gotdotnet
[3] but they are quite minor.

Henrik Frystyk Nielsen
mailto:henrikn@microsoft.com

[1] http://www.ietf.org/internet-drafts/draft-nielsen-dime-00.txt
[2] http://discuss.develop.com/dime.html
[3] http://www.gotdotnet.com/team/xml_wsspecs/dime/default.htm

>-----Original Message-----
>From: James Clark [mailto:jjc@jclark.com] 
>Sent: Friday, November 16, 2001 21:36
>To: Gavin Thomas Nicol; xml-dev@lists.xml.org
>Subject: Re: [xml-dev] Packaging (was Re: [xml-dev] Interoperability)
>
>
>
>--On 16 November 2001 23:42 -0500 Gavin Thomas Nicol 
><gtn@rbii.com> wrote:
>
>> On Friday 16 November 2001 07:44 pm, James Clark wrote:
>>> Another format that should be considered is DIME:
>>>
>>>   http://www.gotdotnet.com/team/xml_wsspecs/dime/default.htm
>>>
>>> It seems pleasantly simple and well-designed. You can stream it for 
>>> both input and output.  It is being used by the Web 
>Services Routing 
>>> Protocol
>>
>> OTOH. I have developed a severe aversion to any MIME-base format for 
>> general  packaging of XML... lot's of issues that I don't 
>even want to 
>> think about  anymore.
>
>In what sense does DIME have a "MIME-base"?  The only 
>connection between 
>DIME and MIME that I can see is that DIME allows you to 
>specify the type of 
>a particular member with a MIME media-type; I can't see 
>anything wrong with 
>that, especially since it gives you a choice of whether you 
>label the type 
>of a member with a MIME media-type or with an absolute URI.
>
>> Compared to XAR (or DZIP) I would say that DIME is more complex...
>
>I'm surprised you think that.  ZIP provides a lot more 
>functionality than 
>DIME:
>
>- file names
>- file attributes
>- compression
>- checksums
>- random access
>
>All DIME does is allow to you pack a sequence of objects into 
>a single byte 
>stream, where each object consists of:
>
>- a type (either a MIME media-type or an absolute URI)
>- a unique identifier (a URI)
>- a sequence of bytes
>
>It does this in a way that is simple and efficient for both 
>reading and 
>writing.