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] Validation vs performance - was Re: [xml-dev] Fast text o

[ Lists Home | Date Index | Thread Index ]

I've nothing against RELAX NG except marketability and that is 
the same for anything new and different.   I don't see evidence 
for 'unstoppable momentum'.  I see programmers getting a 
a sharp knife where before they had to use an axe.  It is a good 
choice of tools for those smart enough to know how to choose. Like 
anything else sold by political, religious or emotional persuasion, 
it's a bad idea to take a knife to an axe fight. :-)

1.  Yes, readability and verbosity are related.  So is comprehension. 
But as you note, it is the complexity of the application or as I have 
noted, the opaqueness of the text that one is measuring, not only the 
syntax.  Even in SGML where end-tag omission was allowed, I found 
that I preferred fully-end tagged text combined with pretty print 
formatting.  So toss in not only syntax but layout.   In the 
simpler C-like (say curlies) syntaxes, I found I had to have a 
brace-sensitive editor regardless.  With XML, I can open it in 
IE and instantly find out if it is well-formed.  With XML I can 
write a DTD, or Schema, or RELAX and find out a lot more.  Cheap 
ubiquitous tools are persuasive if not always sharp.  With every 
other format, I need yetMoreToolsForTheSameJobs.

2.  What value for a measure would be powerful enough to introduce 
a new syntax for an old information set?  One would have to compare 
an unnoticeable cognitive load to a noticeable change in format.

3.  Programmers aren't the only people who have to read it. 

We can learn to read anything we have to.  We can handle alternatives 
if we have to.  We need to know if we have to because optimizing 
for the programmer in all cases is not necessarily a good selector 
for optimizations, and the gains may not be worth the cost.  Again, 
we need to see some 10 for one improvements to consider it.

And again, is a binary characterization WG a mandate to overhaul 
other parts of the system 'while we're at it'.  Scope.  Mission 
creep.  All the usual paranoias.


From: Henrik Martensson [mailto:henrik.martensson@bostream.nu]

Readability has a lot to do with verbosity, and the RELAX NG compact
syntax is certainly a lot less verbose than the XML syntax.

Another point in favor of the compact syntax is the use of regex-like
constructs. As a programmer, this is something I am already comfortable

One could adapt a readability formula, such as Flesch-Kincaid or SMOG to
work with schema syntax. (Well, for the purposes of this discussion,
anyway.) If one did, one would find that the compact syntax is
considerably easier to read, especially for more complex expressions.

Whether one perceives this difference in readability is another matter.
I certainly do. On the other hand, if one is good enough at interpreting
the XML syntax, the extra cognitive load would not be noticeable. It
would probably still be _measurable_ though.

Does anyone know if there has been any work done on the readability of
programming languages? I suspect it would be quite interesting reading.


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

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