Re: [Jsdsi-devel] JSDSI xml

[Sameer A. <aj...@cs...>]

Well, my hope was to avoid requiring any semantic information in the
JSDSI-to-XML translation.  The output I showed you earlier simply converts
jsdsi.Objs to Sexps, then converts these Sexps to XML.  Thus, all the
semantics is captures in the JSDSI-to-Sexp translation.

Using "local-name" as you have below requires the XML tranlation to know
the meaning of a particular string.  I would instead suggest that the Sexp
format be changed to:
(name (public-key ...) (local-name alice) (local-name bob))

Then, everything will work perfectly :)
Sameer

> akin to 'the spec' I would vote for something like the later:
> (name (public-key ...) alice bob)
> becomes
> <name>
>   <public-key>...</public-key>
>   <local-name>alice</local-name>
>   <local-name>bob</local-name>
> </name>
>
> with maybe an (optional) index attribute on the local-name:
> <name>
>   <public-key>...</public-key>
>   <local-name index="0">alice</local-name>
>   <local-name index="1">bob</local-name>
> </name>
>
>
> Sean
>
> On Thu, 2004-02-26 at 13:05, Sameer Ajmani wrote:
>> > Hmm... Those possible changes to the SExp just seem to make them
>> more verbose just for XML...
>> Right.  The alternative is to wrap each base SexpString in its own
>> element: CarolJones
>> becomes
>> <string val="CarolJones"/>
>> or <string>CarolJones</string>
>>
>> > Have you seen this:
>> I glanced at these specs some time age.  We should aim to support this
>> spec whatever the SPKI working group (or Carl Ellison) uses.
>>
>> > http://bladesys.demon.co.uk/roller/resources/sradford/draft-orri-spki-xml-cert-struc-00.txt
>>
>> Thanks,
>> Sameer
>>
>> >
>> > Sean
>> >
>> > On Wed, 2004-02-25 at 18:41, Sameer Ajmani wrote:
>> >> > (at first I wondering if the placing of the 'local name' of
>> >> CarolJones was correct, but I think it is the right thing to do -
>> not many xml documents have a tag and a text value as direct
>> >> children of another tag).
>> >>
>> >> Yes, the string interspersed with elements are ugly (and maybe a
>> problem).
>> >>  I'll try to write an XmlReader that inverts this process sometime
>> >> soon,
>> >> and hopefully that will complain if this is invalid XML :)  It also
>> may be problematic if XML concatenates adjacent strings, e.g.,
>> (name (public-key ...) mary friends) will become:
>> >> <name>
>> >>   <public-key>...</public-key>
>> >>   mary friends
>> >> </name>
>> >>
>> >> XML will read "mary friends" as a single block of character data,
>> while SPKI treasts it as two SexpStrings.  My XmlReader will handle
>> this just fine, but the XML form lacks some structural information.
>> >>
>> >> An alternative is to make SPKI's syntax more XML-compatible by
>> restricting its S-expressions to SexpLists thatcontain one
>> SexpString or an arbitrary number of SexpLists.  So the above name
>> object would be:
>> >> (name (public-key ...) (id mary) (id friends))
>> >>
>> >> While this is ugly in the "name" case, it actually makes some other
>> structures clearer:
>> >>
>> >> (hash md5 |hfslk782yf2|)
>> >> becomes:
>> >> (hash (algo md5) (value |hfslk782yf2|))
>> >>
>> >> (k-of-n "2" "3"
>> >>  (public-key ...) (public-key ...) (public-key ...))
>> >> becomes:
>> >> (k-of-n (k "2") (n "3")
>> >>  (public-key ...) (public-key ...) (public-key ...))
>> >>
>> >> This change would require consulting with the SPKI designers (I
>> could contact Carl Ellison about this).
>> >>
>> >> Sameer
>> >>
>> >> > On Wed, 2004-02-25 at 02:54, Sameer Ajmani wrote:
>> >> >> I have just committed a simple Sexp-to-XML converter to JSDSI.
>> >> This makes it possible to express JSDSI objects as XML (by first
>> converting them to S-expressions).  I added the package jsdsi.xml
>> and the class jsdsi.xml.XmlWriter, and extended the jsdsi.S2X
>> utility to support xml output.  Here's an example:
>> >> >>
>> >> >> Input S-expression:
>> >> >>
>> >> >> (cert
>> >> >>   (issuer
>> >> >>     (name
>> >> >>       (public-key
>> >> >>         (rsa-pkcs1-md5
>> >> >>           (e #010001#)
>> >> >>           (n
>> >> >>             |AKd6zpHoNSq53rFFZ6XZV5m+YZKr1Ym9fUMuE8yHMOZx9AsrWqkZyMxZ9Cn
>> >> >> 5mIPaTLqYEakoJ7lir25a/Nq1IpvFC8mlalpI2LINDRnIt3HbmiFEj97Eaqk
>> 2ffhwXotFHIbPGyrgrMobOtWIpG56Ru5AK230lG68+ZyntCqB53v5|)))
>> >> >>       CarolJones))
>> >> >>   (subject
>> >> >>     (public-key
>> >> >>       (rsa-pkcs1-md5
>> >> >>         (e #010001#)
>> >> >>         (n
>> >> >>           |AK0aiccf7Yj/Gzx93m6R6suJiPr5o9ZON6dG31wwDyGK2aKoMJUYLxVMWGh5R
>> >> >> 2Wsc+RO+ylvablNhmO7zGiZ9qNiN+vdTjdxMALx4ULFF1aYNi86QsSgU/x8Xrk
>> >> 2SiHQ9EkHGDJ7Bpng+4LNSjQtm2DuEscHP3mqz8BzGB+jTek/|)))))
>> >> >>
>> >> >> Output XML:
>> >> >>
>> >> >> <cert>
>> >> >>   <issuer>
>> >> >>     <name>
>> >> >>       <public-key>
>> >> >>         <rsa-pkcs1-md5>
>> >> >>           <e>
>> >> >>             #010001#
>> >> >>           </e>
>> >> >>           <n>
>> >> >>             |AKd6zpHoNSq53rFFZ6XZV5m+YZKr1Ym9fUMuE8yHMOZx9AsrWqkZyMxZ9Cn5
>> >> >> mIPaTLqYEakoJ7lir25a/Nq1IpvFC8mlalpI2LINDRnIt3HbmiFEj97Eaqk2
>> ffhwXotFHIbPGyrgrMobOtWIpG56Ru5AK230lG68+ZyntCqB53v5|
>> >> >>           </n>
>> >> >>         </rsa-pkcs1-md5>
>> >> >>       </public-key>
>> >> >>       CarolJones
>> >> >>     </name>
>> >> >>   </issuer>
>> >> >>   <subject>
>> >> >>     <public-key>
>> >> >>       <rsa-pkcs1-md5>
>> >> >>         <e>
>> >> >>           #010001#
>> >> >>         </e>
>> >> >>         <n>
>> >> >>           |AK0aiccf7Yj/Gzx93m6R6suJiPr5o9ZON6dG31wwDyGK2aKoMJUYLxVMWGh5R2
>> >> >> Wsc+RO+ylvablNhmO7zGiZ9qNiN+vdTjdxMALx4ULFF1aYNi86QsSgU/x8Xrk2
>> >> SiHQ9EkHGDJ7Bpng+4LNSjQtm2DuEscHP3mqz8BzGB+jTek/|
>> >> >>         </n>
>> >> >>       </rsa-pkcs1-md5>
>> >> >>     </public-key>
>> >> >>   </subject>
>> >> >> </cert>
>> >> >>
>> >> >> The problems with this are that it's not terribly compact (e.g.,
>> >> the "e" elements could be all on one line), I'm not sure whether
>> all the characters used in the text blocks are allowed, there's no
>> schema or DTD, and there's no XmlReader (yet).  But does this
>> conversion look reasonable?
>> >> >>
>> >> >> Sameer
>> >> >>
>> >> >> http://ajmani.net
>> >> >>
>> >> >>
>> >> >>
>> >> >>
>> >> >>
>> >> >> -------------------------------------------------------
>> >> >> SF.Net is sponsored by: Speed Start Your Linux Apps Now.
>> >> >> Build and deploy apps & Web services for Linux with
>> >> >> a free DVD software kit from IBM. Click Now!
>> >> >> http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click
>> >> >> _______________________________________________
>> >> >> Jsdsi-devel mailing list
>> >> >> Jsd...@li...
>> >> >> https://lists.sourceforge.net/lists/listinfo/jsdsi-devel
>> >> > --
>> >> > Dr. Sean Radford, MBBS, MSc
>> >> > sra...@ae...
>> >> > http://www.aegeus-technology.com
>> >>
>> >>
>> >> http://ajmani.net
>> >>
>> >>
>> >>
>> >>
>> >>
>> >> -------------------------------------------------------
>> >> SF.Net is sponsored by: Speed Start Your Linux Apps Now.
>> >> Build and deploy apps & Web services for Linux with
>> >> a free DVD software kit from IBM. Click Now!
>> >> http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click
>> >> _______________________________________________
>> >> Jsdsi-devel mailing list
>> >> Jsd...@li...
>> >> https://lists.sourceforge.net/lists/listinfo/jsdsi-devel
>> > --
>> > Dr. Sean Radford, MBBS, MSc
>> > sra...@ae...
>> > http://www.aegeus-technology.com
>>
>>
>> http://ajmani.net
>>
> --
> Dr. Sean Radford, MBBS, MSc
> sra...@ae...
> http://www.aegeus-technology.com


http://ajmani.net