Re:What has TT done for KDE?
on
Qt for Mac
·
· Score: 2
... and there are several other reasonable C++ toolkits.
Honest question: like what ? The only other I know is wxWindows, which is based on GTK+(or Motif) for Unices.
In particular, which one were around, ready for use, when KDE started (like you said in your other post)?
Qt may have been a good initial catalyst and motivator for KDE, but I think the value of KDE greatly overshadows the value of Qt by now.
That is what I call a good investment. We should not complain about all these failing open-source companies , and then complain about the fews that, playing fair, manage to succeed.
To convince me to program fow Windows 2000, you need to prove me that:
Extremely detailed docs about how it works are available, free or at regular publishing costs, possibly from a a variety of sources;
That none of the tools I have to use treats me as mentally impaired ("thrust me, it just works" is good for end-users, maybe, not for developers);
That its design follows the philosophy "not more complex than the problem it wants to solve", and that I can follow the same concept in developing my software ( that is, that I _decide_ to use some of those pretty-acronyms-technologies because I find them useful, not because the tools I use cannot work without them ).
Potential users just don't know what Linux has to offer them. If they will listen, I find that a few minutes of
education and evaluation of their current usage is enough to get them to try Linux.
To try Linux, yes.But I doubt it is enough to use Linux as their primary OS (or only one- double booting is a geek thing which many normal PC users find strange and cumbersome).
Things which prevent people from accepting Linux are:
Hardware support: you can't go in a computer shop, buy something and be sure it works with Linux ( and I have read that also penguin-stickered hardware sometime fails);
Applications: some areas are both not interesting enough to be developed by volunteers and not rewarding enough (given Linux small market) to be developed by companies.
M$oft is used everywere: just like Samba did on the server side, Linux users need better tools to interoperate with M$oft users (first of all a really really good filter for M$oft office docs).
In the seven years I have been a Linux user, progress has been certainly made on all these problems. But, IMHO, not enough. Yet.
That is, a similar idea of Netscape/AOL embedding their bookmarks in their version of Mozilla ( I don't believe this is effective, also, but then,
I don't believe much in advertisement, too ).
Sticking to the point, my reson for being puzzled at Eazel idea is that, being Nautilus open-source, someone else can easily offer the same service (unless they planned to patent the protocol).
Then, the advantage of Eazel would have been only that Nautilus would come pre-configured with their web address (or whatever). And Sun/HP/etc could still struck deals with Eazel competitors and change this, too.
People do not care about which OS their PC is running.
Therefore, the way for Linux to 'dominate the world' is to become invisible. To smootly running boxes of every type without letting people know that it is there ( much like what is it doing now in web servers, print servers, etc... ).
People using this nokia box will not see Linux (at least, this is the hope).They will see the Loki games, plus whatever they put as general GUI (if any).
Hackers around the world will see Linux, however, and then will start playing with it. Some of them may come with nifty new games or software toys (or new tools to build them), which will encourage people to buy more Nokia boxes (again, this is the hope).
I tried but failed to understand which kind of net-based services need Nautilus as an interface.
Wouldn't a web interface ( with add-on infrastructure like Zope) be a better choice?
Also, I never understood the idea 'if they like Nautilus they will come to us for services'. Supposing they were successful, they would have to face:
Companies offering the same service for nautilus-equipped users;
Companies offering similar service with different interface, for people without Nautilus.
This sort of thing only works for closed-source software -- until the protocol is reverse-engineered that is; after it is just lawsuit nightmare ( and lawyers paradise ).
What is the answer to VB in Unix platform. One of the linux gurus had a reply: Python + Tk. While this may be a little far fetched, this is definitely the best available alternative till another one comes by.
There are now. This is my preferred (YMMV):
glade+python(with gtk and libglade modules).
Yo go like this:
build visually yor GUI with glade, a very nice tool for the job.
Instead of generationg code, which IMHO is rather messy, you use libglade to glob the GUI
files and generate on-the-flight all the widgets
you need writing only few lines of code for
file ( it even does 'automatic' callback bindings,
which is great).
You code the logic behind the GUI in python,
using python-gtk to manipulate GUI widgets.
Programming in python might produce better code, but it is enjoyable to write?
Yes. You can have all the fun you want (and more).
Like many Very High Level Languages, Python frees
you from most of the housekeeping a programmer has to do with C, C++ and even Java.
Tanks to its coerent semanthics and clean syntax, you wont have to squeeze your brain trying to
remember obscure rules (or checking brackets)).
Did you ever give-up adding feature X to your program, because it would take too long? Just think : how much of this 'too long' was housekeeping? With python, you might find that
you have the time...
The unix family tree is rather scaring for whoever
would like to support a commercial product on the
whole family... only the powerful magic of autoconf/automake (i.e. sweat and blood of many hackers) can (partially) cope with the miriads of tiny incpmpatibilities among unices.
You're right, possibly.
However, if we get rid definitively of buffer overflows, it will be possible to spend more effort on securing against more tricky vilnerabilities.
I believe that, for instance, a kernel option which makes the stack not executable except by granting a EXECUTE_STACK capability would not hurt linux users, while not being the security panacea that some might expect.
This meant that you weren't even allowed to run KDE at your company without paying a Qt license.
Uhm. IIRC, they always granted free usage of binaries linked with their libraries, and only wanted to be payed if you developed non-free software using their toolkit.
Which, IMHO, is a better revenue model than Eazel asking(begging?) for money.
Said so, I agree that the article is basically FUD which cunningly insists around some not-so-closed sores of the free sofware world ( and tries to open new ones ).
I remember a long thread in debian-user ML just debating this ( a few months ago ).
Someone argued that there is not need to crack the go-gnome.com server : anybody in the middle between you and the server could exploit it to do whatever they want on your machine ( BUT: what would they gain at cracking a workstation, possibly connected with 56K dial-up ?).
OTOH, also typing./configure as root means that you trust the people that prepared the script, as well as all the transmission chain that allowed you to download the binaries ( and how many check a configure script before running it? ).
Sorry (Was Re:How many sinners, then ?)
on
Coder on the Cross
·
· Score: 1
You got modded down.
There is not anymore respect for the eldest!
As far as I know, both KDE and Gnome are quite modular and can be seen as 'Mr. Potato frameworks'. They offere more-or-less-docunmented interface to make your own theme engine, your applets (like pagers), your own Window manager.
I dunno about GNUStep, but I think that it will use another desktop framework, too.
So, the probem is not the lack of a framework, but the lack of a standard framework. Here are working on it, however.
True. No committee.Technical projects must have a lead.
But some process like python's PEP could be a good idea...since every programmer should listen its usersm and the users of computer languages are the other programmers.
I've asked about and can I get a sensible answer about what is the best desktop-come-development-come-nice-windows-style-t hing-but-running-linux distro for me?
A linux distro is more like a toolbox than like finished a product: they are for people that need or like to tweak with things, that are unsatisfied with standard commercial offer.Your needs can be satisfied by any of the mahjor distros, but not without a little work from you.
If you like so much windows, just keep using it. Both Perl and GCC have been ported to it, so you have the tool you need.
Is this pretty much the way everyone feels, or are people more dedicated to a particular distribution?
I can't believe that you are really asking that. This is the mother of all flamebaits:-).
Oh well, maybe the number of hits on slashdot is low, and a flame feast would make CmdTaco & Co more happy (I know, these are difficult times for pernguin-friendly business!).
My idea of a Linux handeld would allow to run the full suit of command-line tools.
Think of:
- mutt for handling mail
- midnight commander for file manager
- bitchx for chatting on irc
- lynx for textual web-browser
- emacs (?!?) for everithing, including preparing your breakfast:-)
If you add to this the multiple virtual consoles,
you may not need any GUI at all
(but only graphic programs to display image files and such).
You will be able to use the 8 MB for something more important that giving your GUI a nice look.
You will need a virtual keyboard, however. And a program like gpm, which allows you to use the pen to interact with course-based or dialog-based user-interface elements.
it seems to be the fashion in the last few years to include all sorts of functionality in every application (modular design isn't an option anymore).
For the little I know about 'component technology' (DCOM or Bonobo or Kparts ), the code is not duplicated. A word-processor can edit drawings because it is 'borrowing' from the drawings editor program not only its funcions but also part of its GUI, wich gets 'embedded' in the word processor.
Said so, I still prefers the old way of sharing code with libraries and starting different processes to perform different activities...
OTOH, the money they pay is _our_ money.
The hard-discount shops base most of their business on selling products which carry little or no ads-induced overprice. And it works - up to a point.
IMO ads are important when you launch a new product (truly new, not like 'our new soap formula will make your linen ever whiter'). Once you have your customer base, they becomes only a reminder, for not loosing customers on new alternative products. But too much reminding can be annoying and it can generates the opposite effect (speaking for myself, there are many products I don't buy because of obnoxious and annoying ads everywhere ).
It would take few minutes to substitute your battery pack with a new loaded one, if there was a recharge station network the size of current gasoline network stations.
Point 1 : half of the sofware you mentioned is not under GPL but have they own licences ( surely python[bsd like], perl (artistic) and apache (can't remember but not GPL )
Point 2 (more important) : One of the clauses in GPL prevents from discriminating software usage for reason of race, religion, ideals etc...
A better idea, IMO, would be embed a new module in next standard distribution of the tools you mentioned : python-decss aka apache-decss aka perl-dcss aka Linux API for DeCSS and so on.
Then letting know Disney and Co that they are using the same 'illegal' software theyr lawyers are trying to ban from the internet.
especially if it is an open source product that most people are expected to compile
I'm afraid given the size of gnome ( as well as of other similar products ), less and less people are willing to compile it from scratch.
And as for all open-source products, the less users try a feature, the more the feature is buggy. Configure scripts can perform their magic flawlessy only because other people ironed out the many bugs in them.
Also the current stable gnome 1.2 reselase has small compiling issues in some package, if you get the tarballs from gnome site.
Slashdot Mirror
Honest question: like what ? The only other I know is wxWindows, which is based on GTK+(or Motif) for Unices.
In particular, which one were around, ready for use, when KDE started (like you said in your other post)?
Qt may have been a good initial catalyst and motivator for KDE, but I think the value of KDE greatly overshadows the value of Qt by now.
That is what I call a good investment. We should not complain about all these failing open-source companies , and then complain about the fews that, playing fair, manage to succeed.To try Linux, yes.But I doubt it is enough to use Linux as their primary OS (or only one- double booting is a geek thing which many normal PC users find strange and cumbersome).
Things which prevent people from accepting Linux are:
- Hardware support: you can't go in a computer shop, buy something and be sure it works with Linux ( and I have read that also penguin-stickered hardware sometime fails);
- Applications: some areas are both not interesting enough to be developed by volunteers and not rewarding enough (given Linux small market) to be developed by companies.
- M$oft is used everywere: just like Samba did on the server side, Linux users need better tools to interoperate with M$oft users (first of all a really really good filter for M$oft office docs).
In the seven years I have been a Linux user, progress has been certainly made on all these problems. But, IMHO, not enough. Yet.Sticking to the point, my reson for being puzzled at Eazel idea is that, being Nautilus open-source, someone else can easily offer the same service (unless they planned to patent the protocol).
Then, the advantage of Eazel would have been only that Nautilus would come pre-configured with their web address (or whatever). And Sun/HP/etc could still struck deals with Eazel competitors and change this, too.
People do not care about which OS their PC is running. Therefore, the way for Linux to 'dominate the world' is to become invisible. To smootly running boxes of every type without letting people know that it is there ( much like what is it doing now in web servers, print servers, etc... ).
People using this nokia box will not see Linux (at least, this is the hope).They will see the Loki games, plus whatever they put as general GUI (if any).
Hackers around the world will see Linux, however, and then will start playing with it. Some of them may come with nifty new games or software toys (or new tools to build them), which will encourage people to buy more Nokia boxes (again, this is the hope).
I guess.
Wouldn't a web interface ( with add-on infrastructure like Zope) be a better choice?
Also, I never understood the idea 'if they like Nautilus they will come to us for services'. Supposing they were successful, they would have to face:
This sort of thing only works for closed-source software -- until the protocol is reverse-engineered that is; after it is just lawsuit nightmare ( and lawyers paradise ).
There are now. This is my preferred (YMMV): glade+python(with gtk and libglade modules). Yo go like this:
Clean and fast.
Yes. You can have all the fun you want (and more). Like many Very High Level Languages, Python frees you from most of the housekeeping a programmer has to do with C, C++ and even Java.
Tanks to its coerent semanthics and clean syntax, you wont have to squeeze your brain trying to remember obscure rules (or checking brackets)).
Did you ever give-up adding feature X to your program, because it would take too long? Just think : how much of this 'too long' was housekeeping? With python, you might find that you have the time ...
The question is: will Linux be able to do better?
However, if we get rid definitively of buffer overflows, it will be possible to spend more effort on securing against more tricky vilnerabilities.
I believe that, for instance, a kernel option which makes the stack not executable except by granting a EXECUTE_STACK capability would not hurt linux users, while not being the security panacea that some might expect.
Uhm. IIRC, they always granted free usage of binaries linked with their libraries, and only wanted to be payed if you developed non-free software using their toolkit.
Which, IMHO, is a better revenue model than Eazel asking(begging?) for money.
Said so, I agree that the article is basically FUD which cunningly insists around some not-so-closed sores of the free sofware world ( and tries to open new ones ).
Someone argued that there is not need to crack the go-gnome.com server : anybody in the middle between you and the server could exploit it to do whatever they want on your machine ( BUT: what would they gain at cracking a workstation, possibly connected with 56K dial-up ?).
OTOH, also typing ./configure as root means that you trust the people that prepared the script, as well as all the transmission chain that allowed you to download the binaries ( and how many check a configure script before running it? ).
You got modded down. There is not anymore respect for the eldest!
I dunno about GNUStep, but I think that it will use another desktop framework, too.
So, the probem is not the lack of a framework, but the lack of a standard framework. Here are working on it, however.
Next time I get a CS degree, I'll be sure to attend some theology class ... just in case.
But some process like python's PEP could be a good idea...since every programmer should listen its usersm and the users of computer languages are the other programmers.
A linux distro is more like a toolbox than like finished a product: they are for people that need or like to tweak with things, that are unsatisfied with standard commercial offer.Your needs can be satisfied by any of the mahjor distros, but not without a little work from you.
If you like so much windows, just keep using it. Both Perl and GCC have been ported to it, so you have the tool you need.
I can't believe that you are really asking that. This is the mother of all flamebaits :-).
Oh well, maybe the number of hits on slashdot is low, and a flame feast would make CmdTaco & Co more happy (I know, these are difficult times for pernguin-friendly business!).
Think of: :-)
- mutt for handling mail
- midnight commander for file manager
- bitchx for chatting on irc - lynx for textual web-browser
- emacs (?!?) for everithing, including preparing your breakfast
If you add to this the multiple virtual consoles, you may not need any GUI at all (but only graphic programs to display image files and such). You will be able to use the 8 MB for something more important that giving your GUI a nice look.
You will need a virtual keyboard, however. And a program like gpm, which allows you to use the pen to interact with course-based or dialog-based user-interface elements.
For the little I know about 'component technology' (DCOM or Bonobo or Kparts ), the code is not duplicated. A word-processor can edit drawings because it is 'borrowing' from the drawings editor program not only its funcions but also part of its GUI, wich gets 'embedded' in the word processor.
Said so, I still prefers the old way of sharing code with libraries and starting different processes to perform different activities ...
The hard-discount shops base most of their business on selling products which carry little or no ads-induced overprice. And it works - up to a point.
IMO ads are important when you launch a new product (truly new, not like 'our new soap formula will make your linen ever whiter'). Once you have your customer base, they becomes only a reminder, for not loosing customers on new alternative products. But too much reminding can be annoying and it can generates the opposite effect (speaking for myself, there are many products I don't buy because of obnoxious and annoying ads everywhere ).
It would take few minutes to substitute your battery pack with a new loaded one, if there was a recharge station network the size of current gasoline network stations.
Point 2 (more important) : One of the clauses in GPL prevents from discriminating software usage for reason of race, religion, ideals etc ...
A better idea, IMO, would be embed a new module in next standard distribution of the tools you mentioned : python-decss aka apache-decss aka perl-dcss aka Linux API for DeCSS and so on.
Then letting know Disney and Co that they are using the same 'illegal' software theyr lawyers are trying to ban from the internet.
I stll have it somewhere. I wonder if I could port Debian on it ... an apt-get miracle --target=vic20 could do the trick, I guess.
I'm afraid given the size of gnome ( as well as of other similar products ), less and less people are willing to compile it from scratch.
And as for all open-source products, the less users try a feature, the more the feature is buggy. Configure scripts can perform their magic flawlessy only because other people ironed out the many bugs in them.
Also the current stable gnome 1.2 reselase has small compiling issues in some package, if you get the tarballs from gnome site.