home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!sun4nl!tuegate.tue.nl!svin09!wsintt05!wsinda
- From: wsinda@wsintt05.info.win.tue.nl (Dick Alstein)
- Newsgroups: comp.sys.acorn.tech
- Subject: Re: New Template format discussion document
- Message-ID: <4926@svin09.info.win.tue.nl>
- Date: 7 Jan 93 14:22:48 GMT
- References: <930101200448@kernow.demon.co.uk> <4897@svin09.info.win.tue.nl> <1993Jan6.113946.4153@gate.esat.kuleuven.ac.be>
- Sender: news@svin09.info.win.tue.nl
- Reply-To: wsinda@info.win.tue.nl
- Lines: 39
-
- tytgat@esat.kuleuven.ac.be (John Tytgat) writes:
-
- >In article <4897@svin09.info.win.tue.nl> wsinda@info.win.tue.nl writes:
- >>* Grouping icons
- >Very good idea. But maybe we can have a look at the way Draw stores grouped
- >objects. I think we can adopt the same way, something like : [...]
- >
- >As in Draw, group of icons can be nested. However, there are certain
- >consequences if we are using this approach : the icon numbering could be the
- >order of apperance in the list of Icon object definitions. And this puts a
- >restriction on grouping/ungrouping because this action can change the order
- >of the icon object definition block.
- >This restriction does not apply for the programmer because (s)he only needs
- >to remake the #define file but that's no solution for the user.
-
- Doing icon grouping as in !Draw has the advantage that it enforces a proper
- nesting (if two groups have objects in common, then one group must be a part of
- the other). However, in Draw files, objects are self-contained (except for
- font handles, I think); all data needed to render an object can be found inside
- the definition block. IMO, it would be best to split the Indirected Data tables
- and attach the data for each icon to the icon definition block.
-
- A consequence is that you are limited in how icons overlap: if icon A overlaps
- icon B, then all icons in the same group as A overlap B. But I think that's not
- a big restriction.
-
- Another (logical) consequence is that you have to hide the icon numbers from
- the programmer; if the grouping changes, then the order of the icon definitions
- in the file may change completely; hence icon numbers are very "unstable", and
- the programmer should only refer to icons by their names.
-
- >What about defining the menu format in the Glass file ?
-
- Looking at Olaf Krumnow's proposal, it wouldn't be much of a problem to define
- it. But, as Tim Browse wrote in his original posting, it's best to start with
- a minimal definition, and extend the format when someone is seriously thinking
- of writing a program (i.e. a Glass editor) that actually uses the extension.
-
- Dick
-