Fixing Things
One of the features in the latest version of ubuntu is a really simple idea for installing software. You just make a link, like this one and when the user clicks on it, they get a prompt offering to install the package, if they say yes then it gets their password and installs it without any fuss.
Because it is a new feature, and it is pretty simple to understand, some people have gone a bit crazy with it - which is cool. But, a usability problem comes up when you have no idea what it is you are actually installing:
It is difficult to answer this question meaninfully. My thought process would be that it uses qt (which is a GUI library), has a gui (well I already suspected that from the qt part), and does something with "pfs", which according to google is "Professional Freight Services". But since all the packages in Ubuntu have a description associated with them, we can just add an extra bit to the interface:
So we can find out what it is.
Adding such a feature is quite simple, and you'd hope that with my computer science degree and years of experience that it wouldn't be much of an issue. So, I downloaded the code for it, added the new bits and then generated a file which says what changes I made, so then someone can integrate it into the main codebase that everyone uses.
Unfortunately, I tend to get really stupid when it comes to the final step of getting things integrated. My understanding of things is that every patch has a 'bug' that goes with it. So for a patch to be accepted, it needs to be fixing something. So I went to the place that handles all the bugs and found this:
There are currently no open bugs.
Now I don't think for an instant that a newly added feature will have no open bugs, so I look around and find on the overview page that it "Doesn’t use Bugs". So obviously the bugs are tracked somewhere else. Because I get stupid here, I have no idea where to look.
This happened yesterday, so today I thought that I'd give it another shot. So I google for the bug page again, and find this:
On a page that looks otherwise very similar to the one yesterday (which is also similar to this one). There I found that someone had already had this idea, so I put my patch in with that bug, and I guess if it gets accepted then we'll see it in ubuntu later.
The reason all of this is interesting to me is that I think contributing needs to be really easy - even if I'm really dumb. I've been scared off reporting bugs for some software packages before because I'm faced with screens like this:
I don't mean to specifically bag out one group. But, I can't imagine many people being able to correctly answer the question about what subcomponent they are talking about.
But it is a bit of a balancing act. The developers want the bugs neatly categorised by the users, and the user just wants a textbox where they can rant about whatever. I don't think I've dealt with a system where you can report bugs without logging in, but I think that is important. If someone is reporting a bug, I think you need to make it as easy as possible for them, and then have slightly more knowledgable people come along and do the categorisation. The problem is that that is a sucky and boring job for most people, so it is a bit of a tough sell.
I think in the case of the form above, the improvements I would make would be:
- Remove the requirement for having an account
- Give every field an 'unsure' option
- By default, hide the fields other than the summary and the description
- Add an (optional) email address field for a non-registered user, so they can stay in the loop
I think it is a lot easier to categorise a vague bug description than to work out what a user was about to post, than to guess what the bug was that a user was going to post, but then didn't because they couldn't be bothered to figure out what hoops they need to jump through.
RSS