Re: User Facing Documents

Bijan
    I don't think any of us are talking end users in the sense of  
"grandma" but if we're going to see more OWL uptake, many of us  
believe that people will want at least some understanding of the  
language "while they stand on one foot" to decide whether to dig in  
deeper.  Further, we've seen with OWL 1.0 that there are some  
communities that are willing to pick and choose a few OWL constructs  
to use in their applications.  I don't think that is a bad thing, and  
it provides an "on ramp" to OWL.  FOAF, SKOS and vocabularies in the  
individual sciences are starting to pick up on some OWL constructs,  
and I think it is good to keep this coming.
   Most of these people are not trained logicians or people with DL  
backgrounds, and given how impenetrable I find the current three  
documents, and I do have that training, I'm guessing a lot of them  
simply won't take the time to see if the new constructs are of use,  
until there's a simple way to look at them and get a feel.  I think  
it is important that we let people know what OWL may look like after  
the new version, and see what kind of response we get from the  
application designers using OWL, rather than only from the people  
developing OWL "power tools" like Pellet and SWOOP  [1]
  it's also clear to me that we will eventually have to create a  
reference for the new features, as well as some sort of examples that  
people can cut and paste to make sure they get the syntax right while  
learning the language
    So while you are right that end-user studies are very tied to  
particular users, that is not what any of us seem to be proposing.   
Rather, we're looklng for an easier way than the current documents  
for people to "Grok the fullness" of OWL 1.1's new features and to  
give us feedback (as well as to get excited about those things in the  
new version they might see as useful - which will also be necessary  
later in the process when we are trying to get the AC excited)
   -JH



[1] truth in advertising - Pellet was originally developed in my lab  
at Maryland, it's development through the first three versions was  
completely funded by grants which I was the PI on, and I have  
coauthored many papers relating to its use and work - so I am not  
anti-Pellet (and its ilk) by any means.  (Ditto for SWOOP - now the  
second most used OWL editor after Protege...)


On Oct 31, 2007, at 8:03 AM, Bijan Parsia wrote:

>
> [I trust I can post this without getting accused of being anti- 
> user, user-unknowing, narrowly technical, or whatever other  
> dismissal de jour is in the air.]
>
> It is a truism of HCI that users vary, so if you are going to  
> evaluate something it has to be with respect to specific users (or  
> user types) and specific issues. So, it would be helpful if we  
> could identify a bit better what users various documents are  
> targeting. It's just not *feasible* to reach all the users we might  
> want to reach, in the best way for them.
>
> For example, are we going to develop tutorials? I don't think  
> that's a good use of WG time and effort. *Recognizing* and  
> marketing and encouraging third party tutorials is sensible.  
> Tweaking documents in small ways to accommodated tutorial writers  
> is *quite* sensible (because we get a lot of bang for the buck).
>
> The two most important things we can do for users, which are most  
> squarely in our mandate and solidly within our power, is to produce  
> a language that maximally useful for them with specifications that  
> support -- nay, encourage -- a strong, actively developed  
> infrastructure (of tools, expertise, experience, community, and  
> success).
>
> It's also important that user facing websites (which is really what  
> we are talking about) are actively maintain ad infinitum. Ephemera  
> should be clearly labeled as such (best if naturally so like blog  
> posts). Pace contrevening evidence e.g., from server logs, I'd  
> imagine that a "What is OWL?" document on xml.com:
>
> 	<http://xml.com/>
>
> is probably more valuable overall. It targets an influential market  
> (i.e., web developers with some technical savvy) in a focused way.  
> Plus, it could helps elide, a little, the perception that OWL is  
> niche, or W3C top down wankery.
>
> Similarly, at OWLED 2007, a participant called not for more "basic"  
> user documentation, but for "mid level" stuff, i.e., you've kinda  
> learned the language, now what? E.g., things more similar to the  
> SWBPD patterns stuff. Hence, the OWLED task force on "Education":
>
> 	<http://code.google.com/p/owl1-1/wiki/Education>
>
> Things like implementation lists:
>
> 	<http://code.google.com/p/owl1-1/wiki/Implementations>
> 	<http://www.cs.man.ac.uk/~sattler/reasoners.html>
>
> are really useful, but generally only if maintained. So, for many  
> classes of user facing document, we need a post-WG sustainability  
> strategy.
>
> OWLED and webont.org are two venues for this. I urge people not to  
> forget that the WG is only one way to work toward the success of  
> OWL, and that it is an instrument with limits and particular  
> strengths.
>
> For example, adding a workable keys proposal to the language is, in  
> my estimation, far more important than adding *any* user facing  
> document. I base this on the fact that the users and "customers"  
> that Matthew, Robert, and Alan influence don't seem to read any of  
> the documents. They just ask Matthew, Robert, and Alan. They tell  
> me that the lack of easy peasy keys is a huge barrier to adoption,  
> or even to starting conversations about adoption. Extending the  
> language is something the WG can, almost uniquely, do.
>
> (And note that speccing easy keys has not been trivial. Uli and I  
> have spent more time that we expected trying to nail down all the  
> details. And of course, no one cares about a *mere* spec...they  
> need them *deployed*, which I am working on and constrained the  
> specification.)
>
> Cheers,
> Bijan.
>

"If we knew what we were doing, it wouldn't be called research, would  
it?." - Albert Einstein

Prof James Hendler				http://www.cs.rpi.edu/~hendler
Tetherless World Constellation Chair
Computer Science Dept
Rensselaer Polytechnic Institute, Troy NY 12180

Received on Wednesday, 31 October 2007 14:47:20 UTC