rdl-devel — Discussion of development related issues

Re: [rdl-devel] RDL Object Model

[R.S. S. <rss...@ya...>]

--- Efi Somos <efi...@ho...> wrote:
> So who's doing the object model?

Dave had volunteered on that one. I see I forgot to update
the task assignment. Will do that.

__________________________________
Do you Yahoo!?
Yahoo! Hotjobs: Enter the "Signing Bonus" Sweepstakes
http://hotjobs.sweepstakes.yahoo.com/signingbonus

[rdl-devel] RDL Object Model

[Efi S. <efi...@ho...>]

So who's doing the object model?
 
Efi

Re: [rdl-devel] Little Update

[R.S. S. <rss...@ya...>]

--- Dave Gill <dav...@bl...> wrote:
> 
> I have 'want' in my environment but it needs some work
> doing on it to
> make it Linux friendly. 

I asked about this on the want mailing list. In the mean
time, do we have an alternative? The thing is, I don't care
much for make.


> The nagging
> thing is that the Libraries of source code it relies on,
> after a cursory
> glance, look windows specific.  
> Not sure how much time I
> should spend on
> that to be honest. I know the task is assigned to me so
> do you want me
> to hack want into Linux shape?

I took a quick look, and it seems to me that the only thing
that is Win32 specific is the Crt32 unit. Am I right?

This should be relatively easy to port using curses? Or
maybe even just using plain WriteLns to get it running
quickly? Do you want (no pun intended) to spend a little
time on it?


> I also had a look at some of the source code in CVS and
> my only
> immediate comment would be around variable naming. Can we
> use some
> prefix for the variable names to show scope and type? 

I'm not a big fan of Hungarian notation. The idea of having
to rename my variable when the type changes, yuck.

Knowing the scope can be helpfull. For this, I use the
following simple scheme:
- local variables start with a lowercase letter
- arguments to a routine start with an lowercase 'a'
followed by an uppercase letter
- properties start with an uppercase letter
- module level variables (globals defined in the
implementation) start with 'MODULE_'
- global variables (which I hope we won't be using) start
with 'GLOBAL_'

What do you think?


> I know for
> Instantiated objects that type is a bit of a nonsense but
> we could
> prefix it with a 'c' so in long lines of code we know
> what we are dealing with.

I prefer not to have long lines of code ;-) Really, when
the code becomes obscure enough that these things start to
matter, I think you are way beyond the point where
refactoring for clarity is wanted.


> Do we have an example RDL document
> to look at? 

Yes, it's in /example.


__________________________________
Do you Yahoo!?
Yahoo! Hotjobs: Enter the "Signing Bonus" Sweepstakes
http://hotjobs.sweepstakes.yahoo.com/signingbonus

[rdl-devel] Little Update

[Dave G. <dav...@bl...>]

Hi Guys,

Latest bits of information.

Yes GNU/Linux will be my main area of work. Simply because I have been
coding in Windows for years and need the challenge of getting all this
working in GNU/Linux. I currently have Redhat 9 on a 2.4 Kernel.

I have 'want' in my environment but it needs some work doing on it to
make it Linux friendly. For example there is a Batch file to run to do
an initial build of it. This uses the resource compiler which I do not
have in Linux. Well I have an equivalent which is a time expiry
shareware equivalent. So the batch file would need converting to a bash
script and then the whole thing would need to be checked. The nagging
thing is that the Libraries of source code it relies on, after a cursory
glance, look windows specific.  Not sure how much time I should spend on
that to be honest. I know the task is assigned to me so do you want me
to hack want into Linux shape?

I also had a look at some of the source code in CVS and my only
immediate comment would be around variable naming. Can we use some
prefix for the variable names to show scope and type? I know for
Instantiated objects that type is a bit of a nonsense but we could
prefix it with a 'c' so in long lines of code we know what we are
dealing with.

Composite pattern for the RDL object model seems reasonable. There are a
few questions that may need to be asked around implementation of that
but that is for later. Do we have an example RDL document to look at? So
I can make this abstract model in my head a little more substantial.

I think that's it for now.

Dave.

 

 

[rdl-devel] test

[<rsi...@pl...>]