Comments on James Clark's Random Thoughts: Some thoughts on the Oslo Modeling Language

@James:I think I may have submitted my last commen...

2008-12-10T10:14:00.000+07:00

@James:

I think I may have submitted my last comment more than once, and possibly more than twice. "How?" you might ask?

Yeah, you got me on that one, cuz' I'm not really sure. But there was definitely at least one "did that get submitted correctly?" statement followed by a "I don't think it did" followed by a "Oh /PLEASE/ tell me I didn't just resubmit it *TWICE*!".

You'd think after programming these damn things for the last 25 years I would have got the hang of how to use them.

Apparently not. ;-)

Anyway, please feel free to delete however many duplicates made their way through, and in the mean time:

You: "Absolutely. Entering the standards process now would be very premature. Design by committee tends not to produce very elegant results. But I would like standardization to be on the agenda eventually, when the time is ripe (which I would have thought would be several years away at least)."

Can't help but agree. For the Record, and FWIW, it seems to me that "M" is something that the "Big Dogs" at MSFT have been planning for quite some time. See: http://www.xsltblog.com/archives/2005/11/is_the_agile_de.html for a better understanding of what I am referring to (closer to the bottom where I mention Chris Sells "LISP" comment is where things sorta come together...)

In other news, what reminded me about this post was Doug Purdy's post [http://douglaspurdy.com/2008/12/05/james-clark-on-oslo-m/] from a few days ago which, at present time, is linked to on the front page of the MSDN "Oslo" front page under the "Oslo Team Blogs" header in the bottom left hand corner of the page: http://msdn.microsoft.com/en-us/oslo/default.aspx

"Big Brother" is paying attention, apparently. Good for Big Brother! It's about time he started realizing that little brother knows twice as much, is twice as fast, and is half his size.

The word "agile" comes to mind. ;-)

Go show Big Brother what little brother is all about, James! w00t! :-D

@Kingsley:Conceptually speaking what's the differe...

2008-12-10T09:48:00.000+07:00

@Kingsley:

Conceptually speaking what's the difference between what C is trying to accomplish and what LISP is trying to accomplish? Scratching your head at why on earth C was ever needed in the first place doesn't change the fact that C hit a developers sweet spot that to this day LISP has yet to hit.

Is LISP a better overall language than C? I think so, but I also look at XSLT and see poetry while others see utter chaos and confusion.

RDF and M might theoretically solve the same domain of problem(s) just as LISP and C might do the same for their respective domain(s).

But C won. Why?

Arguing that one way of solving a problem simply mirrors an existing way of solving the /same/ problem does not mean the argument has been won. Instead, it means that if you find yourself arguing why the predecessor is better than the follower, it's probably because the follower now owns 90% of the same market the predecessor is lucky to still be consider a contender in.

Worser is better, Kingsley. Just ask Bill Gates and Steve Jobs. ;-)

@Kingsley:Conceptually speaking what's the differe...

2008-12-10T09:47:00.000+07:00

What Kingsley Idehen said. Is there any indication...

2008-12-08T23:05:00.000+07:00

What Kingsley Idehen said.

Is there any indication that MS is targeting XML with this language or is it possible they're out to snip the very gradual growth of RDF before it becomes serious competition? Or, put another way, maybe M could revitalize the Semantic Web push? RDF has had years of slow growth but so far it's kind of floundered about. A serious commitment by MS to an alternate syntax might but just what the SW needs.

James,How is "M" conceptually distinct from RDF an...

2008-12-02T23:35:00.000+07:00

James,

How is "M" conceptually distinct from RDF and the Semantic Web vision in general? In both cases we are looking at Entity / Conceptual Level Data Modelling.

I continue to scratch my head about the degrees of cognitive dissonace exhibited by these efforts :-)

Kingsley Idehen

Thanks for taking the time to post this review of ...

2008-11-25T03:39:00.000+07:00

Thanks for taking the time to post this review of M, James! It was just the kind of summary I was looking for.
Cheers, Tony.

Absolutely. Entering the standards process now wou...

2008-11-24T10:03:00.000+07:00

Absolutely. Entering the standards process now would be very premature. Design by committee tends not to produce very elegant results. But I would like standardization to be on the agenda eventually, when the time is ripe (which I would have thought would be several years away at least).

@callingshotgun:>> This seems similar to Google's ...

2008-11-24T09:20:00.000+07:00

@callingshotgun:

>> This seems similar to Google's open-sourced "Protocol Buffers." How do they compare in terms of pros/cons? <<

Huh? Protocol Buffers are a binary data serialization format. Suggesting ProtoBuf's and the M language represent the same thing would be like suggesting,

'(foo '(bar baz))

... and ...

(foo '(bar baz))

... represent the same thing.

For the record, they don't. ProtoBuf's are binary data structures. Like the first sample from above, PB's just represent data. Nothing more, nothing less.

On the other hand, like LISP/Scheme, M combines both code and data such as that represented by the second sample.

Another example: XSLT == XML, but XML certainly != XSLT.

>> I can understand this: if I was Microsoft, I c...

2008-11-24T08:58:00.000+07:00

>> I can understand this: if I was Microsoft, I certainly wouldn't be keen to repeat the XSD or OOXML experience. But open source is not a substitute for standardization. <<

True, but isn't making an attempt to standardize a language that exists in an alpha state a bit premature? After reading your analysis I'm /REALLY/ liking what I'm seeing: It's like the best of Scheme, XSLT, and XML wrapped into a C-style syntax that's feels familiar to a broader swath of folks.

Let's start talking about standardization once a field-tested 1.0 release hits the streets. In the mean time, fuck standardization: Let creativity and real-world problem solving take its due course. Then standardize once the language feels complete enough to start imposing "You MUST do this to be considered a standards compliant implementation."

Or am I missing something more obvious that suggests letting the standardization process reign supreme over the creativity process is a better approach than just letting creative, brilliant people such as yourself get involved with the creation process before then taking things to the next obvious level of standardization.

I hope not, because /your/ brain and talent involved with the creation of a wrist friendly executable version of XSLT causes feelings of "Oh, Dear God... What could the collective forces of James Clark, Don Box, Chris Sells, and other prominent members of the "If I had my way, S-Expressions inside of LISP/Scheme/DSSSL/related dialects would reign supreme" community create if they weren't forced to work through the confines of a standardization body until they were God damn ready to work with a standardization body to bring an official definition to the end result of their open sourced creative process of creating a language that couldn't be criticized by anyone with the credentials for their criticisms to actually matter.

Hell, I'd bet even Knuth would find joy in the end result of just such a creation from just such a group of folks.

James: Get involved. PLEASE!

Looks like JSON to me!!

2008-11-23T23:38:00.000+07:00

Looks like JSON to me!!

Interesting post- Hadn't looked into Oslo yet. A...

2008-11-23T22:31:00.000+07:00

Interesting post- Hadn't looked into Oslo yet. A question, though- This seems similar to Google's open-sourced "Protocol Buffers." How do they compare in terms of pros/cons?