On 05/04/11 21:35, Devayon Das wrote:
> Hi Jeroen,
> Thanks for clarifying. Could you elaborate on where else such UI
> usability improvements could be made?
> Also, about the SMW query language, I have found documentation on
> http://semantic-mediawiki.org/wiki/Help:Semantic_search and the related
> links. Is there any further documentation available anywhere else?
I think more insights can mainly be found in the code:
The internal structure of query descriptions is defined by the objects
in . While this could use some more documentation, it captures the
very core of what a query condition /is/. The basic types of printout
requests are defined in . Each part of the query processing and
result formatting accepts some parameters. Most relate to query
formatting, and each query printer (see ) documents its own parameters.
This said, I do not at all think that the overall task of building a
replacement for Special:Ask is too small for 3 months. Maybe the
opposite is the case. The query builder in Halo has been constructed by
quite a few people over a long time, and I am still not really convinced
by the interface. What we envision is a really compelling solution,
easy, scalable to user needs, responsive, extendible, and most
importantly: cleanly and elegantly coded. In other words, we are not
talking about improvements to the existing code but about completely new
code to replace it.
The UI design as such will be a challenge, since nobody (AFAICT) has yet
built a UI for such expressive queries for a broad audience. Is it even
possible to achieve something that /all/ readers can use? This would
involve even the people who cannot edit pages. This might require us to
have different types of UI -- we may need more than two tabs for "basic"
and "expert" views! The possible UI approaches are very wide. Some
obvious paradigms include:
* form-based/query centred (current Special:Ask)
* form-based/data centred (like Semantic Forms for queries)
* browsing based (like Drilldown or Exhibit)
* structured query builder (like in Halo)
* graphical query builder (idea: draw up the searched-for relationships
with a WireIt interface to specify a query)
* some combination of the above (e.g. browsing-based with form inputs
for property values)
Obviously, all of these very diffferent views share some common aspects
(same query model, desire to view and browse results, same parameters
for printout, ...). So one would like to have a software architecture
that provides such basics and allows "query UI modules" to be used on
top of it. This should be extendible, i.e. extension should be able to
register query UI classes that can be plugged into Special:Ask. A lot of
architectural challenges here, and certainly a lot of coding to get two
or three basic modules set up.
Finally, all of this should be as interactive as possible while still
being merciful on the server (live updates of query results, for
UIs is necessary, but the basic viewing of "further results" must still
I think that the above is already more than one can hope for in three
months, on average. Within this area, the focus can be put on those
things that are most interesting to you. If you should unexpectedly run
out of work after 2 months, then we will have no problem to find a lot
more high-impact activities in this field to continue (such as
contextual search features accessible from the article view -- "search
for similar pages").
Anyway, don't be overwhelmed by the above. We can go step by step. In
any case, these changes can have an enormous impact in an area where SMW
so far lacks a consistent approach. Most of the changes would directly
affect user experience of hundreds or thousands of users, and maybe even
open a path for other developers to contribute further search UI
extensions (further increasing the practical impact). I am happy to
mentor one or more projects in this space, or co-mentor together with
> On Tue, Apr 5, 2011 at 6:21 PM, Jeroen De Dauw <jeroendedauw@...
> <mailto:jeroendedauw@...>> wrote:
> Hey Dev,
> > I see that many people scoff at UI projects as being to 'easy'?
> Is it a good idea to post as a proposal?
> I think this is an important project for increasing SMW usability,
> and that it has been proposed by Markus is also an indicator it's a
> good idea :) The project might not be _that_ easy, as the project
> description states, you need to be familiar with all the
> capabilities of the SMW query language and be able to create an
> interface that's truly easy to understand and use (while retaining
> all features). My main concern is that this project is probably not
> enough work to cover 3 months. However, if you run out of work there
> are similar improvements that can be made at other places in SMW and
> related extensions.
> I'd be willing to mentor, or preferably co-mentor, this. I'm
> familiar with the involved code and the SMW query language, but am
> not an UI/usability person, so would be of limited help there.
> Jeroen De Dauw
> http://www.bn2vs.com <http://www.bn2vs.com/>
> Don't panic. Don't be evil.
> Xperia(TM) PLAY
> It's a major breakthrough. An authentic gaming
> smartphone on the nation's most reliable network.
> And it wants your games.
> Semediawiki-devel mailing list