Why OPML is winning

Dave Winer points to the anti-OPML questioning of an 'XML geek' and quotes the pro-OPML answer of an 'ordinary' user - "Because there are tools".

As a wannabe programming geek myself (I just never had the aptitude) I can understand, to a degree, what motivates these engineers who are almost pained by the widespread adoption of a 'sub-optimal' specification. But as Raymond says -

"OPML gives me something that xml does not: Tools that I can begin to actually use, and OPML is also a format/specification/language/whatever that is already being used out there."

And that's where the perfectionist geeks need to get a clue. I actually bought a book about XML five years ago with the full intention of 'getting into it'. But its still sitting on the shelf because my ongoing perception was that XML continued to remain obscure and impractical. It was only after I'd been happily using RSS for some time that I realized it was actually an XML spec. "My God, so this XML really is useful for something.... who would have thought!"

What was the difference between RSS and other formats, to me a lowly user? Simply that it was useful for something! And that there were all sorts of wonderful tools emerging daily to take advantage of it. It wasn't some wonderful technology just sitting in a lab - it was practical and beneficial.

And so it is with OPML. Even though most node managers working on the Open Irish Directory had never worked with it before they could immediately grok its significance. And they could build their first OPML node within a few minutes using OPMLmanager.com (or Dave Winer's own OPML Editor). Not only are editors proliferating but so also are OPML browsers/viewers including Taskable, Yabfog, KBCafe and Koz OPML.

I would like the 'XML geeks' to show me what comparable tools there are for any other XML spec?

Comments

[Danny]

While dismissing me as an "XML geek" (which I may well be) Dave also failed to mention my follow up point - that there are lots more tools for (X)HTML than there are for OPML. You can do all the same things in HTML that you can do in OPML - links, lists, hierarchy. On the left hand side of this blog I side a column marked "Categories" - where did that appear from? OPML tools?

Dave's just trying to market a subset of the existing Web done using his own format. Caveat emptor.

[Danny]

PS. Your final question - what comparable tools there are for any other XML spec? Depends what you want to do. If you want to do OPML-like stuff, Google "XHTML tools".

[phil jones]

Danny,

if you do Google XHTML tools, you find things that are much more generic than the application(s) specific OPML editor.

The art of software creation is to discover interesting sweet-spots which capture the syntactic commonality between a lot of different things people want to do, while adding useful constraints that simplify the task, and just enough semantics to make the whole thing plug into the real-world requirements of users.

Classic genres of software : word-processors, spread-sheets, relational databases etc. are all good examples of this. The word processor is a generic symbol editor, that doesn't make a distinction between love-letters and company reports, but has enough semantic knowledge to check the spelling.

The spread-sheet doesn't care what you are modelling, but has enough semantics to know the difference between numbers and words, and how to do mathematical operations.

These are very useful combinations.

And this is the kind of sweet-spot that end-users are looking for in tools. As far as I can see, there really isn't anything else with the profile of the OPML editor : for giving me a generic syntactic commonality (editing nested lists) with enough semantics (eg. turn this into a blog) to be exciting.

[Ken MacLeod]

One of the things you'll find about the geeks is that they're (er, we're) a lot better at talking than doing.

The obvious other format is XOXO (a subset of HTML), but are any of the geeks working with the outliner developers to produce, consume, and edit XOXO? There's no real world use of an alternate format. No oomph to get it going and flesh it out.

[Dave Winer]

Phil, those are good points -- I'd add this.

It's possible to edit a spreadsheet with a text editor. A text editor user might be confused, hearing about a spreadsheet for the first time, saying "I do that all the time with my text editor."

I've heard that about outliners going all the way back to my graduate studies at the University of Wisconsin in the late 70s. Programmers don't know that there are editors can understand structure and give users a way to directly manipulate it.

That's leverage. If I know what you want to do, I can groom the software to do that well. Later we saw the same effect as word processing software tried to add outlining as a feature. They failed, not because they are bad designers or engineers, but because they had to make choices in their user interface, go one way and you get a word processor, go the other you get an outliner. The decisions are mutually exclusive. There's only one set of arrow keys. Double-clicking must do one thing, it can't do the right thing for both applications.

I'm sure the XML geeks are well-intentioned (not sure what else to call them, they didn't like wonks either), but they are having as much trouble accepting this as the programmers did at UW in the 70s.

I wrote a story about this here...

http://davewiner.userland.com/outlinersProgramming

Keep up the discussion, it's very useful.

[Danny]

Phil, I don't have a problem with the outliner application idea, it just bugs me that it's bringing in an unnecessary complication in the form of a new format. Surely the OPML Editor could use XHTML as its format..?

Rewording what you said a little: a HTML editor has generic syntactic commonality (editing nested lists, and more) with enough semantics (eg. this *is* a blog).

Ken, I'd suggest the real-world use is the Web ;-)

Ok, I must admit having a narrower range of capability like OPML compared to XHTML can probably improve the focus of developers. But so far all we're seeing is tools that implement little subsets of what Web tools can already do, in a largely incompatible format. Where's the innovation in that?

I don't actually care very much about formats as long as they're reasonably specified - XSLT can be a marvel for conversion.

[Danny]

Dave, my last post went out of sequence.

"Programmers don't know that there are editors can understand structure and give users a way to directly manipulate it."

Er, yes they do, and have done at least since Emacs appeared (probably earlier). I would guess that the majority of programmers on the planet use IDEs which all allow structural viewing of the code. In fact many current tools (Eclipse, NetBeans, almost certainly Visual Studio) go well into visual, structural manipulation of the semantics of the code, far beyond a simple hierarchical view.

"If I know what you want to do, I can groom the software to do that well."

I agree entirely. Like I said in the last comment, I don't have a problem with the outliner application idea, sure, use the arrow keys for outline expansion. I only wonder why a new format should be necessary when one of the most widely-used formats on the planet (HTML) can do the same thing. This is orthogonal from the user interface.

[Danny]

PS. Here's an in-browser outline editor based on HTML :
http://www.decafbad.com/blog/2005/07/12/xoxo_outliner_experiment

See also:
http://decafbad.com/blog/2005/11/24/do-gophers-make-good-stuffing

[Bill Seitz]

Maybe you guys need to "adopt" an open-source blogging tool and aggregator, and built little plugins for each to work with XHTML/XOXO.

For instance, maybe all a blogroll needs is to be wrapped in a "class" named "blogroll"?

[Dave Winer]

Bill, the OPML Editor is open source. GPL.

I'm going to release an aggregator shortly, that is also GPL.

[David Mercer]

My wordpress trackbacks are either off or busted (can't tell! :-) so I'll just toss the link to my reply here (quite long, and an addition to the point here, which I agree with):

http://davidmercer.nfshost.com/article/4/TheOTHERreasonOPMLissucceedingandrockstheNet

[David Masterson]

Well, obviously, OPML never caught on and became an important format. I can really only find two usages of OPML currently -- the OPML Editor (by Dave Winer) and Carbonfin's Outliner. Last time I looked at the OPML editor, it was somewhat difficult to setup and use (kind of expected for a GPL product). Carbonfin's Outliner only works on iPhone/iPad and may be using its own form of OPML internally. What I've been looking for from Carbonfin is for it's Outliner to implement import/export from Emacs Outline form since Emacs is implemented on most everything else and, even if it isn't, it's Outline form is much simpler and more useful than OPML.