shilling for my feature request [1926]

Mark Brownell gizmotron at earthlink.net
Sat Jul 31 10:03:31 CDT 2004


On Friday, July 30, 2004, at 09:09 PM, Troy Rollins wrote:

> Very interesting. You certainly are quite the evangelist for the 
> merits of pull-parsing. I read the reference docs, but I have to 
> admit, I'm not the parsing method connoisseur I suppose I should be. 
> I've always used the "whatever works" approach. To that end, I've used 
> both DOM and SAX, and rolled-my-own in other instances. I'm still not 
> positive I'd know when to say "this needs a pull-parser!" Nor can I 
> claim to fully understand all the benefits and efficiencies of MTML, 
> fortunately, I doubt I am alone in that.  ;-)
> --
> Troy

Troy,

What I'm proposing in this feature request is not really a complete 
pull-parser. It's the speedy engine behind rapid development of high 
speed custom parsers that can be used to create the object modal type 
pull-parser or the fragment type modal, my invention type. Imagine 
extracting XML records into separate record objects but leaving the 
internal XML markup for each record still intact for later. This would 
be seen as fragments of the fuller XML data.   MTML, as an XML like 
structure, is rule based. It is not validated or requiring well 
formedness to work properly, hence the need for custom parsing.

MTML can be an XML data structure but it can also be an extensible 
markup extension to HTML and for the purpose of gathering common text 
fragments from HTML. For example I have been running experiments on 
various versions of a directory of Gestalt Psychology practitioners 
from around the world. The data is meant to be rendered by its Rev 
style HTML in a field object. So it reads like a web page. Powerful 
search tools allows the user to find specific information from the full 
document in order to find a Psychologist or Therapist in their area 
that handles their specific issues. In a single search they can gather 
all that fit their request and display just the information that they 
were interested in finding. So what you have is a browser window that 
displays text and images, plays video and sound files, and can have 
extensible markup added to it by the user while viewing it. If the user 
wishes to add relational connections that make convenient recollection 
of important information available the use of the human readable 
extensibility is seamlessly added to the background HTML and is 
savable. What really floated my pirate ship was I added the internet 
and created four custom MTML hyperlinks that allows the user to click 
on links to MTML files on the internet and / or open them as the new 
file or merge them with the existing current file. So website 
developers can offer pre-tagged MTML files that have the relational 
text gathering power already built into their presentations.  The other 
two MTML links are keyword searches of traditional websites and MTML 
embedded into the code of an existing HTML webpage. I found out that 
Internet Explorer and Netscape Navigator ignore MTML in the HTML and 
render it just fine.

So in conclusion MTML is not really a substitute for XML but it does 
tend to fill the promise of a simpler XML if so structured that way. 
It's most powerful feature is in the fact that it is meant to be added 
to common readable text as a background tool that is not seen.

Mark



More information about the use-livecode mailing list