-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1

On Wednesday 04 Jun 2003 1:48 pm, linuxdev wrote:

...

Trust me file chooser is a first year programming >project, but at the same time most used dialog. C >and callback mechanism of event handling in UI is a stumbling block, and in absence of OO, things >quickly become bad.

You answered your own question there. And i add that they do have a nice

And by doing this I proved GTK was a bad choice!!

roadmap in place and almost there. They have managed to a hell of a job

Test yuor statement yourself. Find a page corresponding to http://developer.kde.org/development-versions/kde-3.2-features.html for Gnome, and objectively see how good are they doing.

without OO. Kudos to the gnome team for working out things.

Kudos for surviving with such basic design flaws. OO is a good thing not evil. Gnome hopes to, and even falsely promotes their toolkit as object oriented.

Your reference to KDE's slowness was an article of the year 2001 with gcc 2.95 in question. and even there the author points out that the faults lie with kde and not

If you have reasons to believe what they have written on that page esp: <quote> With a properly installed KDE3 and a relatively recent GCC/glibc one can expect start-up times that are on par with systems written entirely in C or with very little library integration. </quote> is incorrect, you can drop a mail to kde lists, or atlease the maintainer. Otherwise please dont spread myth.

gcc. I seriously disagree on the point of RH exploiting GPL. They give you freedom to modify whatever stuff you get from them. So you simply cannot whine.

GNU wont whine if RH wud have done that, they woud have sued it :-) Giving freedom to modify is not RHs option, its their compulsion.

The "demeaning" theory that you talked about only started floating around when rh modified kde (m'be even stripped it a little) for its 8.1 psyche. for 9.0 they give you the full dosh.

It has history in pre 7.0 era too, I dont know specific details on the origin os this "dispute". But having paid employee churning out Gnome, and neglecting KDE stays even today. Against nearly all contemporary distros and linux user surveys, RH uses Gnome as primary desktop. Calling Gnome and KDE comparable is demeaning :-)

...

IceWM! IceWM! Its the coolest one.

??

Google! Google! Its the coolest one :-)

...

Using GTK is the biggest mistake done by gaim

AND by abiword and by openoffice and by few hundred other perfectly useful other apps.

I said it for Gaim, and meant only for it. Others you mentioned are from the era when GTK was more GPL than Qt. Besides Qt is CPP, and if the rest of program is in C I wont really recommend CPP GUI.

BTW are you sure Abiword and OO.o are Gnome apps? AFAIK they are GUI library independent in the core, use GTK in case of linux at final stage for rendering. In cases like this it simply doesnt matter what toolkit you use, its not a GTK/Gnome application.Similer paradigm is used for Mozilla, and it took 6 days for Qt team to port mozilla to Qt as fun exercise [Its not supported anymore], so I dont cracterize them any more GTK than Qt, give or take a few weeks :-)

And guess what ? they separated the core and the gui in the latest release of gaim. The plugin architechture has been cleaned up too.

Umm.. Tell me, can you run the core without GUI? Can you put some other GUI on the core? What is the purpose of any Core/GUI separation? They tried, they failed, and satisfied with a compromise solution. Nothing has been really cleaned up, plugin architecture has changed thats it. Dont believe everything anyone says. What the community was hoping was some core which will lets say let them use it with cron to send message to their mobile and other wierd things [and dont even tell me you dont want such a thing], 6.0 was a great disappointment. They failed because GTK is such a mess.

Gnome simply doesn't do it better than KDE for you because you choose it not to. For me it does :D. The default options of both kde and gnome suck

Default options are not those of KDE or Gnome, they are of distributors. Mandrake's default options for KDE are abs great, Galaxy theme is cool, fonts great. Please dont bring this into discussion again. I was just asking if you have reasons to feel that despite this KDE requires more user tweaking than Gnome as you seemed to have claimed more than once.

Have a nice day,

- -- Amit Upadhyay Senior Undergraduate Student Department of Mechanical Engg. Indian Institute of Technology Bombay Mumbai-76, India Phone: (91) 9820325940

On Wed, 4 Jun 2003 17:22:33 +0530 Amit Upadhyay wrote:

Gnome hopes to, and even falsely promotes their toolkit as object oriented.

Why do you say that? Because Gnome is written in C? One can write object oriented code in C too. The last time I had a look at GTK+ 1.2 docs (ca. 2001), I found the design to be object-oriented in nature.

On Jun 4, 2003 at 17:35, Richard Unitek wrote:

I want to build a multi level tree system using php and database like mysql. e.g./

1
   / | \
  /  |  \
 2	3   4

simple tree structure with 3 Childs and so on.

You need a data structure called a tree (surprise!). You could have a struct like this:

struct foo { nodedata; struct foo **children; /* list of pointers */ }

In perl, you'd have an anonymous array instead of the struct **.

----- Original Message ----- From: "Richard Unitek" richardc@unitek.co.in To: "GNU/Linux Users Group, Mumbai, India" linuxers@mm.ilug-bom.org.in Sent: Wednesday, June 04, 2003 4:05 PM Subject: [ILUG-BOM] [OT] PHP - building/simulating tree structure.

Hi

This is an off topic posting.

I want to build a multi level tree system using php and database like

mysql.

Firstly, note that it's PHP and MySQL. OK, I can see you want to use _databases_ and I assume you want trees to make categories and subcategories.... that support as many as levels you want.

Well, you'd do it this way: 1. Every node/leaf has to be identifiable, so every node/leaf will have to have an ID.

2. Nodes/leaves other than the root will be children of someone... so there must be some mechanism to identify the parent. Use the Parent's ID...

In database talk, you'd use a table with two keys... one primary and a foreign. The foreign key will refer the parent key of it's own table.

There are obviously better ways of doing this, but everything depends on where you want to apply the code. Let me know if you need further help _off-list_.

And oh yes, there is alreay a lot of code that does things like this. One can learn a lot by just looking at existing code rather than posting at lists in panic mode.

Regards,

ah.