> " list of configurations grows faster than the number of facets" 
> I would put this down to abuse of configurations than anything else. 

> Personally to me, it is much better anyday to select a single item in 

> configurations than having to go and select individual

 

The problem is that "abuse" is in the eye of the beholder. What I have seen
is that it is natural for the facet creator to view their facet as having
extra importance and therefore deserving of broader exposure via a preset.
As there is no impartial central arbitrator of what constitutes a
"deserving" preset nor could you define a criteria for such central
arbitration, the list grows quickly due to the cross product effect. As
people's usecases and perceptions differ widely, the argument over the
merits of a particular preset is not one that anyone can ever hope win. 

 

If you need an example of this effect, take a look at the XDoclet facet and
it associated presets.

 

> Personally to me, it is much better anyday to select a single item in 

> configurations than having to go and select individual facets.

 

That is true in principle, but not in reality. Once you have a large list of
presets named like "x with y and z", it takes more time to scan the list to
find the one that you are looking for than it would to select the facets
directly. The problem is even worse for novice users. The present names may
not be obvious. The descriptions often don't help much. At the same time,
the concrete definition of what you are selecting (the actual facets) is
hidden from view. In my observation of both novice and experienced users, I
have seen presets as being more of a hindrance than help. The only people
who I have seen thrive on presets are tooling QA engineers who have to
create the same configurations over and over again.

 

> I just hope we address the right problem.

 

Indeed.

 

- Konstantin

 

 

From: [email protected] [mailto:[email protected]] On
Behalf Of Jaspreet Singh
Sent: Wednesday, December 23, 2009 10:25 PM
To: General discussion of project-wide or architectural issues.
Cc: [email protected]; [email protected]
Subject: Re: [wtp-dev] Improving discoverability of facets in WTP wizards

 


Hi, 

" list of configurations grows faster than the number of facets" 
I would put this down to abuse of configurations than anything else.
Personally to me, it is much better anyday to select a single item in
configurations than having to go and select individual facets. I just hope
we address the right problem. 
We might as well just rename the existing "Modify" button to either "Modify
Technologies..." or just "Technologies...". 






"Konstantin Komissarchik" <[email protected]> 
Sent by: [email protected] 

12/24/2009 02:53 AM 


Please respond to
"General discussion of project-wide or architectural issues."
<[email protected]>


To

<[email protected]> 


cc

        

Subject

[wtp-dev] Improving discoverability of facets in WTP wizards

 

                




I'd like to solicit feedback about some proposed changes. The changes are
tracked by the following bug, but I am also repeating them in this e-mail
for convenience. 
  
 <https://bugs.eclipse.org/bugs/show_bug.cgi?id=298483>
https://bugs.eclipse.org/bugs/show_bug.cgi?id=298483 
  
Facet selection page used to be the second page in all WTP project creation
wizards. A few releases ago it was removed out of the wizard flow and placed
into a separate dialog accessible via the "Edit" button next to the
"Configurations" combo box on the first page of the wizard. 
  
Since then, we have started hearing numerous feedback from novice users that
indicates that we have hidden facets too much. The most recent feedback was
a usability study conducted by a team at Oracle where users who have never
used WTP were tasked with writing a basic application. One of the shocking
discoveries is that they couldn't figure out how to enable JPA. These
weren't novice programmers either (the team was composed of devs from our
WebLogic Server group), but they were new to our tooling. 
  
Time is right to take another look at this problem. I've had a discussion
with a UI designer at Oracle about this and together we have came up with a
mockup (attached) that proposes the following changes that should in
principle make facets more discoverable without doing anything too drastic: 
  
1. Create a "Technologies" group that merges the module spec version frame
and the facet configurations frame. 
  
2. Eliminate the facet configurations drop-down. Configurations (aka
presets) is one of those ideas that I largely consider a failure. They sound
good in principle, but do not actually improve usability as the list of
configurations grows faster than the number of facets due to the
cross-product effect. Once the list is long, picking through it is next to
impossible, especially for novice users who this feature is trying to help
in the first place. This change doesn't propose that configurations are
eliminated completely (although that's something that we should seriously
consider in the future). They would still be available in the facet
selection panel. 
  
3. Convert the edit button into a more obvious "Choose other technologies to
use in the project..." hyperlink. 
  
- Konstantin_______________________________________________
wtp-dev mailing list
[email protected]
https://dev.eclipse.org/mailman/listinfo/wtp-dev

<<image001.gif>>

_______________________________________________
wtp-dev mailing list
[email protected]
https://dev.eclipse.org/mailman/listinfo/wtp-dev

Reply via email to