Mathieu Tozer's Dev Blog

Just a thouht I had in the shower a couple of days ago...

All this debate over whether the future of software is on the web or on the desktop or whatever.... I believe the answer is that it's going to be on both. Because software designers need to step back and realise that it's not about software, it's about user information, and they also need to reassess their view of software to match something more like their views of hardware - generic boxes that house stuff. The attitude should be that software will be a generic housing too, for the really valuable stuff - people's information, and that information should be managed wherever the user is, whether it be on the web or desktop. Integrated and synchronised seamlessly, access to information should happen on whichever box, and whichever software.

I think the two will coexist too, one won't push out the other.

What is words ere networked software and people started to use it to develop their own languages? And kids started to build their own in group talk.

Also words should have some mechanism for adding lists of words with their definitions from people's existing lists.

And some kind of quixksilver-esque way of quickly adding and defining words.

Now that I'm moving on to developing the smart group system, I have to think about how users are going to edit the predicates.
I have to first of all
a) learn about predicates myself
b) Implement it stupidly so that the user has to know something about them as well.
c) implement or insert some kind of prefabricated UI predicate editor which sets the smart list predicates depending on word data.

Let's start with the library.

When a language is created, a smart list called 'library' is also created.
A new word 'Welcome' is associated to the new language.
When the word array is updated by the groupController with the Library smart list selected, it will fetch all words and plop the resulting array into the word array.

Then we should be done.

Hard code the library predicate into the library object, and don't allow users to change it.

I used the smartgroup code from CoreRecipes and it built me first go a class which provides me with a nice accessor method - (NSSet *) words; which will do all the hard work and give me a set to put wherever I like for all the words that are in the smart group. I can also use a similar interface method to set the predicate too.

Might have screwed up something with my manual groups though... some objects aren't lasting! :-(

Anyway it's time for bed.

Groups
The sets aren't persisting.... actually, they were! I just wasn't telling my code that a ManualGroup was a manualGroup, and the program was skipping right over the display update code.

Nice to use the debugger to fix this and then find some old word objects I thought I had lost forever attached to some old manual groups :)

Another bug to iron out.

2006-08-01 11:59:09.912 Groups[10213] Filling the word array with words from this manual group
2006-08-01 11:59:10.091 Groups[10213] Exception raised during posting of notification. Ignored. exception: Cannot perform operation without a managed object context

Who is calling this. WHAT doesn't have access to a moc?
Do we ever really know how what we've created works?

Anyway, it seems my update isn't working properly, But it looks as though objects are being added properly and managed in the right way. My MODEL is correct, but my methods for the VIEW is still struggling somewhat.

Hmm. I know.

WELL GET A MANAGED OBJECT CONTEXT!!!!!!


I've got word objects associating themselves to manual groups, and being managed by their selection.
They stay in the application from launch to launch (they persist) and all is well and dandy.

Moving right along.

It was ASKING for a managedObjectContext, so I gave it one, right within interface builder :D
Seems to update itself with each click now. Will it persist....? Hey hey hey!!! It all works.

All shiny and sparky...

Manuaaaaaal Groups!!!!

Just finished V2 of this document for the sake of formalising what I want to achieve in the final project this semester.
You can read it in PDF format. It will be there as long as I am a Monash Student (it's hosted on their server).

-Completed Words Specification Version 2.0, which is a stripped version of the original document which was unnecessarily lengthy.
-Continued work on Groups, mistakenly believing that I was almost done with it.
-Effectively used the debugger to find errors in my code.
-Added a Managed Object Context in Interface Builder which eventually turned out to be in error, but solved a problem at the time.
-Manual grouping system worked with word objects persisting in the store.
-Identified the challenges and things I would need to know about to implement smart groups.
-While attempting implementation of smart groups broke manual groups, indicating that perhaps the design is not as modular as might be desirable.
-Had an 'ideas day' where several new and high level feature ideas for Words came about.