- From: Jacob Palme <jpalme@dsv.su.se>
- Date: Thu, 2 Mar 2000 18:10:53 +0100
- To: General discussion of Extensible Markup Language <XML-L@LISTSERV.HEANET.IE>, IETF Applications Area Discussion List <discuss@apps.ietf.org>
I have made a short comparison of ABNF, ASN.1 and XML, by encoding the same information with all three coding methods. The result of the comparison is not very surprising. The meta description (ABNF, ASN.1 resp. DTD) was about equally long with all three languages. My personal preference is for ASN.1 as a meta language, since it is easy to read and gives stronger type control. The encoded data required 169 octets for ABNF, 54 octets for ASN.1-BER and 297 octets for XML. ASN.1-PER would probably give even less octets. The readability of the encoded data was however much much better with ABNF and XML. XML gives especially good readability of the encoded data for complex information structures. And the need for many more octets with ABNF and especially XML can of course be eliminated through compression. I made some tests with a compression program, and it is able to eliminate almost all the additional redundancy with ABNF and XML, but maybe not for very small files. My personal wish would be to have such a neat and strongly typed metalanguage as ASN.1, but such neat and readable encoded data as XML. (I know that there is something called XER. But I do not believe that is what I am asking for. XER is an attempt to encode ASN.1 into a subset of XML. What I would wish is an ASN.1-like replacement for the DTD meta language, i.e. a meta language which can replace DTD for the full span of the XML language.) This message makes a number of statements on very controversial issues, so all of you may not agree with the conclusions! The comparison can be found at: http://dsv.su.se/jpalme/abook/comparison.pdf in Acrobat format, and at http://dsv.su.se/jpalme/abook/comparison.ps in Postscript format (Note: My BER encoding is probably not 100% right. It is difficult to produce correct BER encoding manually, and I have no ASN.1-BER encoding program to help me. But this will probably not influence the results of the comparison very much.) -- Jacob Palme <jpalme@dsv.su.se> (Stockholm University and KTH) for more info see URL: http://www.dsv.su.se/jpalme/
Received on Thursday, 2 March 2000 12:14:22 UTC