Canola decisions
So, I was reading Murray’s review on Canola and after that reading also comments mainly from some guys in our Canola team, mainly trying to justify some of the things Murray thinks are without reasong. Simply ui design errors. As he was not the first one to talk about that, and of course our team is small, but has different background people, etc and even the interns talked about some of the things he complains, I kind of wanted to “talk” a little bit about the 2 more “controversial” issues in Canola: Web configure and not being able to select any item.
I will use his comment, but I could have used some other 4 or 5 guys comments as well:
So there is a separate Canola Configuration menu item, which starts a web page which connects to a local web server. There’s no reason for this to be a web UI, and no reason for it to be a separate application, interrupting the user experience.
Funny. Since the beginning we said that Canola would be simple, so our main goal is try to fulfill the common end user requirements(use with remote control, hardware keys, and no “movable selection”) and for that some sacrifices needed to be made. One of those was to focus on Stability, Speed, and the basic features working. So what kind makes me think is that Murray just said : “there’s no reason for this to be a web ui” where in my opinion he should at least have asked us “why did you guys did this damn web ui?” =)
If the developers want to get and set application configuration via a local http service then that’s fine
Again, we didn’t used the “client-server” paradigm thinking of that and then created a web ui for that. We did the web server because of the UI restrictions and developers resources to do so. Was the other way around.
The same goes for radio stations and podcasts. Please let me choose them from a list or wizard via the regular UI. At the moment they are even hidden under a “plugins” section in the web UI. Give me instant gratification now, please.
Well, I agree we should have a set of “default” rss / podcast enabled and also that the information architecture in the Config is crap. But again, we have the time constraint. This interaction with picking up good podcasts, ranking them, asking permission to include in the build, was not defined, and trying to “push” it in the end of development time would be a hell. This is however not an excuse, it’s a fact. A lot of things could be done better. But simply there’s was no time AND resources to do that. Imagine typing a RSS feed URL on Canola, UI even on the Thumbpad. That’s not user experience. The basic experience for that : is copy and pasting. It would not be possible inside Canola as it is today.
Those who were involved since the beginning knows that I’m always listening to people, and even more : I don’t criticize features, don’t say “just no” to them. I ponder, measure, rank, all of them and with those things I’m kind of able to decided what fits and what doesn’t in our roadmap for Canola. If you want to give a try in the tracker for example, sometimes we placed there our doubts and “problems”. So I do again the same here that unfortunately killed our “Configuration /Settings” items on the main menu:
- How to enable input. This is the basic. How to enable users to input data, without breaking the user experience?
We discussed a lot about it, and the only way was to develop a “Canola keyboard”. Remember then when we’re deciding this there’s was no thumbpad or whatsoever, and the 770 SDL application using GTK components is pretty complicated. But how much time would we need to create a usable, supporting international languages keyboard? As the UI designer I had a couple of ideas how to implement simple input, but for complete / usable input it was not enough. - With input plugged in, how much time would be needed to develop the widgets needed. Ok, I fully agree with you in the Theme change, but come on : I would place a new icon in the menu, just to change themes. The “configuration” icon was present in the menu until we reached the point of : there’s no resources to make a fully-canola-like configure screen inside Canola.
The point that I’m trying to reach here is : most of the basic Canola features available, can be used without the configuration tool, and that was our goal. It’s just a way to provide more control, without compromising the user experience. But something we in the team all agree:
It needs a lot of more work, to be at least good.
But the important question is:
Can we create the configuration inside Canola?
Yes we can and will do it. But SDL is not the wise way to go. That’s why our plans changed, and the boys are working hard with Evas. Evas will give us the flexibility we need to improve the UI. Not only with configuration, but also features that are out due to limitations in SDL.
I was writing about, Canola future, and I can at least repeat here : The only sure thing about it is that it will be simple. I know people will complain about X or Y, and of course we are listening, and improving it to fit most of the needs, but come on, there’s a limit and if we need to choose again between simplicity (less features, 2 ui) or complexity(1 ui) I would stay with the first one. I want to reach the point where the team will be able to do the 1 ui only, but we are unfortunately not there.
Others can do it? For sure! That’s the beauty of software development. Anyone with skills and 1 pc can do it. And we are not worried with that, we are just trying to create what we believe: simple software, easy to use, very stable and that regular users can really understand what’s happening. If it was for geeks only, I would make a lot of crazy things in the ui. Specially if I was the user in mind, 2 thumbs shortcuts, split scroll, gestures and motion direction detection and a lot of other crazy crazy things would be there!
Murray, this is nothing personal. I hope you don’t take it as so. It’s absolute the opposite. It’s to show that we are really listening, and that we are not blind, but decisions were made and to change them we needed a real rework, that is now happening =)
Thanks for the insight. Things always seem simpler than they really are, and people (like me) will always demand more. But what you have done achieved is already very usable, and you deserve a lot of credit for having a strong focus.
Of course, your time constraints will be helped when you r bosses eventually allow you to open source Canola. I look forward to that eagerly.
Thanks.
March 11th, 2007 at 6:48 amThe web UI for the configuration is in fact a problem, especially since you can *not* connect to it unless you are somehow connected to an actual wireless basestation or using a mobile phone to get a network connection.
The browser simply does not allow to connect to localhost unless you have a real network connection going and I don’t happen to have access to a basestation wherever I am. In fact, I tend to use the Nokia 770 in offline mode quite a lot.
March 11th, 2007 at 10:41 amHi,
I have just started to use Conola and I am impressed.
Just a couple of suggestions.
1. Could all media apps agree on where to keep their covers? So that as I move between them they don’t go through the downloading again. ( This is a little petty as once you have canola why would you need another media player? It could be done using the .thumbnails like the file managers do with symlinks for tracks from the same album, still it would be nice if it could be agreed at fd.o)
2. Could you put an option in the canola interface that takes you to the config web window. I don’t mind the web interface as an idea keeps the main interface nice and clean. It is a little slow at the moment but how often should you need to change the settings?
3. With the new plugins in the browser would it be possible to add a menu option to “Add stream to canola” when you push-and-hold a link.
Thanks again for a nice app.
September 12th, 2007 at 6:31 amAbout RSS / podcast subscription — how about registering a Canola helper to be the handler for rss/xml mime types? That way (with MicroB at least) the user can browse to the website for a podcast, click on the link and get the podcast added to Canola.
By the way, could not comment from Firefox/Linux-i686 just now (failed hash check). Bizarre.
September 13th, 2007 at 9:24 am