IMO financial applications are the number one sample why the Linux desktop hasn't gotten more than a few percents market share. Can you really imagine an ordinary dentist ever use GnuCash on his Windows box? Or a carpenter, or a house wife? Can you really think that such people go out, buy or install Linux on their computers? No that won't happen, not until the art of writing FOSS has changed dramatically.
That said financial FOSS applications will only become possible when they are true cross-platform, when they are available on Windows and MacOS as well. Yet that's not sufficient, they also have to look native and they have to feel native on any platform. Else people, who use computer as tools and not as gadgets, won't use them.
Ordinary people don't look with the eyes of a fan, the look with the eyes of an annoyed worker who wants it task done as fast as possible. None of the so far mentioned applications look acceptable in their eyes. At the current state none written in Java or with GTK will satisfy these people. The only choice which produces acceptable results are using the commercial QT or the free wxWidgets toolkit. It may sound harsh but that's the case, just listen to the complains these people bring up against FOSS applications (or read http://www.osdl.org/dtl/DTL_Survey_Report_Nov2005. pdf).
Yet looking acceptable is only one step towards broad acceptance, the other step is feel acceptable. Sorry, a FOSS application following the Gnome UI guidelines does not feel acceptable on Windows, MacOSX, KDE, etc. If you port a Gnome application to another platform you have to take care of all the little details which are different, which annoy users when the don't fit. These little details are listed in the only cross-platform guidelines wyoGuide (http://wyoguide.sf.net/).
To summarize, to make a FOSS financial application successful you have to follow these simple steps:
- Design the application cross-platform, then you get enough market share. - Write it with an acceptable look anywhere, use QT or wxWidgets. - Care for acceptable feel anywhere, follow the cross-platform guidelines wyoGuide.
You said "try wxWidgets". I did.I summarize my results:
Windows: Ok results. Application built with a wxWidgets builder.
Does this mean you built your own sample with RAD tool like wxDesigner or CodeBlocks? Then your test is more about RAD tools than wxWidgets. BTW why didn't you use wyoGuide-Demo as your starting code base?
wxWidgets on Linux is based on GTK+, therefore any wxWidgets should look equally nice as any other GTK+ application. If it doesn't you're doing something wrong. Yet I've suggested several times to wxWidgets developer to create a DirectFB (GTKDFB) port which would allow for much nicer Linux applications.
wxWidgets on Windows is based on Win32, therefore looks as nice as any MFC application and far better than any GTK+ application.
I don't know wxWidgets on Solaris (I've never tried it) but the developer capacity is rather limited. Complain to Sun if you aren't happy.
Total time preparing a "cross platform" wxWidgets application: 75 hours. But then, I *may* not know what I am doing.
Just compare once these 75 hours for 4 platforms with the one year Google needed to move GoogleEarth to Linux. Sure a full featured application may need a few hundred hours but then you're set up throughout the full development cycle.
Everytime Slashdot mentions Linux, you bring up wyoGuide...
That's not right, I'll bring wyoGuide up whenever it fits, so you won't find any mentioning in the Linux kernel discussions.
Please send feedback directly to the wyoguide-users mailing list, the chance of getting lost or forgotten is much smaller. Of course I try to not forgetting your comment here.
Many people, as also I, don't know much of OSDL beside Linus Torvalds is there employed or they care for Carrier Grade Linux (whatever that means). Yet I know OSDL has done a survey about why the Linux desktop isn't a success (http://www.osdl.org/dtl/DTL_Survey_Report_Nov2005 .pdf).
But now what? Even if the reasons now are more than obvious does the OSDL take the next needed steps? Sure OSDL has created the Portland initiative, unfortunately these people aren't able to do anything about the most pressing matter, the first top inhibitor for the Linux desktop adoption. It might be these people simply don't know how to fix this problem albeit I've shown them one possible solution (http://lists.osdl.org/pipermail/desktop_architect s/2005-December/000349.html).
OSDL might say that "they represent the Linux community", yet OSDL isn't able to bring Linux to success, to increase its market share to a significant amount. So I would think twice if to participate in such an organization. It's sad when even the self proclaimed speaker of the Linux community can't do better.
To say it once more, without agreeing on a single set of application guidelines, guidelines which enhance the usability and the look&feel, there's no hope. All one can say is "Yet another year without a Linux desktop".
I think you will find that quote is usually attributed to Albert Einstein...
Yet Prof. Wirth used it all the time even if he wasn't original author. I wouldn't be surprised if it isn't Einstein either. So while people get known to use a quote that never implies they've invented it.
I'm using the term "schitter bis bewölkt" (sorry German) on all possible occasions and if I ever get as famous as Einstein, people would certainly attribute it to me albeit I definitely can't reclaim authorship.;-)
Why should anybody create a distributed application when a simple library API is almost always sufficient. Why making something more complex than necessary. In most cases component APIs are rather stable as soon as all the missing pieces from early beta versions are solved. Yet even if they change it's possible to handle most cases without any intermediate interface definition etc.
Prof. Wirth always said: "Make everything as simple as possible but not simpler". While Prof. Wirth as made many things too simple (to prove his statement true) any component system is yet much too complex for any locale task and many times also even for distributed tasks. I'm still waiting for a component system which is as easy usable as a simple library API.
don't write code which needs several month to get accustomed! That's the same mistake as is usually done in science. This isn't limited to science or computers, it's all over the world.
Yet computer code is more sensitive to this kind of miss interpretation since neider coders nor their bosses know all the necessary "things" making their code more understandable for others. That's one important reason I still stick to the sample code within the wyoGuide guidelines (http://wyoguide.sf.net/) even if it gets rated biased towards wxWidgets. IMO it's incredible important for OpenSource as for ClosedSource to make code as readable as possible and nothing is better suited to show how than sample code.
Sure enough it would make a lot sense to have sample code for other frameworks or environments but sorry, as a spare time developer I don't have that much time.
Ok, so call me an un-trendy old-fart. Let's see..... a quick review of my resume since
1980, when I first became a programmer.....
I wrote my first Fortran program in 1974 and it work fine on an IBM computer and a few years later on a Digital PDP-11. Albeit it was my first cross-platform code, there's no way to sell it these days.
... that languages are not, nor have they ever been, nor will they EVER be, a magic bullet.
Languages were never the problem and even for the same task several different languages could be used. But you can't expect users to enter countless numbers on a console or copy back the resulting numbers. So even if you code a simple square root function you have to code a GUI around which is many times more code than the function itself. Yet if you want to sell your function depending on its kind users expects sufficiently good look&feel.
Actually my wife has worked with Linux for about a year with Linux in a dual-boot environment because her computer wasn't fast enough under Windows. But after she got a new one she switched back to Windows-only because of the following reasons:
My wife uses exclusively the following applications: - A browser (Mozilla/SeaMonkey) - An office (MSOffice, Excel/Word) - A mailer (Outlook) - A music recorder/editor (Audacity)
The reason my wife used the dual-boot Linux was, Audacity under Windows couldn't then record music with a sufficient acceptable quality. After this limitation was lifted she dropped Linux in favor of Windows because there isn't a mailer like Outlook on Linux. She didn't object using OpenOffice albeit the 1.1 version wasn't a real match to MSOffice at that time. You might wonder why my wife insisted on using Outlook. Simply at office she is pressed to use Outlook and didn't like using a second mailer.
To make Linux more attractive for my wife I don't suggest to build yet an Outlook clone on Linux albeit this would help. Yet the real solution would be building a cross-platform mailer which outperforms Outlook on Windows. Outperforms not in the eyes of a Linux-Fan but in the eyes of an ordinary business worker doing his job.
Yes. Slashdot has quite a reputation for attracting knowledgeable people, yet be aware that some are rather biased towards OpenSource. And don't forget that people who do OpenSource (me including) have a rather absolute opinion. So as long as you are a little sceptic you should be able extract the trends.
As you mention Java you may well notice that currently any Java discussion always tends toward flame wars. Flame wars are always signs that something isn't good, that the there isn't an uphill trend. Flame wars always arises when the future (a trend) isn't going as wished.
I'm probably much biased but IMO the future trend in software development is "cross-platform". So far for many years you could do resonable cross-platform development only with Java. Today you can equally well do cross-plaform development with AJAX or with wyoGuide (binary applications, http://wyoguide.sf.net/). So regardless which of the different technology takes the lead, cross-platform development will increase to the point where single-platform development won't be accepted.
Yep I've seen this too. It's never a good thing to just take something away without giving an alternative.
Beside as soon as people starts involving the IT, it's inportant enough so they'll pay for it. If not, it's not worth the effort. We have made very good experience with applying a price tag to such kind of stuff.
While for developers code may count it doesn't for users. If native code dies is only a questions if developers are able to build easily native code as if none-native code. As long as there are useful tools around. An these days this can be safely answered with yes.
On the other side users care at lot for native GUIs. As long as users are able to distinct e.g. a Java application from a binary (native code) application within 5 min. of using, vendors selling SW have to care for native code which produces far better native GUIs than none-native code.
Yet the SW industry does not work as in other branches. Since development costs is the driving force other considerations have to be taken into account. So for inhouse development, where users complains don't matter that much, none-native Java is favored and forced. On the other side in the free world and the commercial vendors there's no alternative to native GUIs therefore native code.
Yet the future of native versus not native isn't decided so far. With the hyped AJAX there's a none-native technique on the rise while with wxWidgets/wyoGuide (http://wyoguide.sf.net/) there's a now superb cross-platform native technique available. So it might be that in the end there's a draw between native and none-native going along very well.
I'm amazed, even the title of this story is wrong. For browsing any tool who can display code is sufficient. Sure tools which can syntax highlight or have name reference lookup are better but the real issue is to find the source code first hand.
Whenever I search for a solution I first go to http://koders.com/ but their index seems a little limited. Still just try once looking up "wxSingleInstanceChecker". There are others like Koders.com but I've forgotten their names. Next I try to think of a most fitting statement for code and feed it to Google (e.g. "class App: public wxApp", but beware of the white space) which at least returns some hints which project might have source code available. Yet the free text search isn't very suited for searching code since it produces too much wrong results.
When I've found a project which might have fitting code I either look into their LXR if it's available or simply download the source tarbal and use a decent editor (e.g. http://freshmeat.net/projects/wyoeditor/).
It's said that none of the current CVS web tools are searchable, nor that Google is able to restrict results to CVS pages, else it would be much easier to search for source code.
What I meant, was exactly which aspect of the Linux marketplace is Debian "leading" in ?
You seem to be not a developer else you would know what "leading" I meant. For many years OpenSource developers preferred Debian over all other distributions because of various reasons. IMO the most important reason is, you could easily mix a stable distribution with experimental packages which sometimes were essential for development. Yet Debian has fallen back in providing enough recent packages and yet Ubuntu can equally be intermixed. So Debian is loosing upstream developers. I might overestimate or you might underestimate what this effect means for Debian but it's definitely not good.
It's sad that Debian always provoke such ideologic wars as if no more important work and decisions are needed. What does it matter if Java is in none-free or else. Who cares much except users who just want to download a distribution as easy as possible.
IMO there are bigger issues at stake for Debian. The upcoming Ubuntu distribution is threatening to replace Debian as the leading distribution. The GPL v3 does nowhere fit in well into the DFSG. The Linux desktop while each year announced does not take place. It seems Debian people just fights these wars so nobody notices the real threads Debian is faced with.
I've myself switched to Ubuntu since I'm not happy with Debian anymore and I'm most probably don't write about the Linux desktop in the Debian mailing lists anymore etc. Any I'm quite sure quite a lot of others do the same. I've simply got tired of all this ideologic stuff.
I would argue not using VB on the basis that it is not cross platform....
I second that, if you become successful with your product sooner or later you'll be confronted with this question. As soon as you want to sell your product to governments or city administrations you'll be faced with this question.
Yet VB definitely has no long time future and you have to switch anytime. If you surely know you always stay in the Windows environment you might choose C# but I wouldn't advice that. You can equally well choose to switch to C++/MFC (as somebody else advised) and still have the road open to anything. From C++/MFC it's quite easy to switch to C++/wxWidgets which you should target in any case in the long run. With wxWidgets your code almost instantly becomes cross-platform without much extra work. All you have to do is keep an eye on wyoGuide (http://wyoguide.sf.net/), it advises you how to code so your product becomes cross-platform even if you never actually try it out.
So before you switch design your development strategy first. Since development always needs time, sometimes years, and you have to support your product even longer. So IMO moving slowly towards C++/wxWidgets pays off in any case. Any good developer knows C++ and wxWidgets is quite a full featured framework.
There are more ways to develop good products but I simply can't advise them. Java with all its addins isn't easier to code than C++ anymore, faster development time has become a joke and Java applications are easily spotted by the users and definitely not liked. The QT framework is another alternative but there seems to be no free cross-platform applications and when you look how long it takes to port GoogleEarth to Linux I'm loosing fait. Yet the GTK+ framework, forgive me if I say it so blunt, might be useful for free SW but in no case for commercial development.
It might be that the LSB makes life easier for distribution but does it also have an effect for developers and users? I don't remember I ever have looked into the LSB when designing and coding an application nor when distributing source files. And I'm quite sure most users don't even know that the LSB exist. While the LSB is very important for the binary distribution, it's influence on a Linux system is rather limited. Yet the FSG only cares for the LSB and therefore it's importance is also rather limited.
A useful Linux system needs some user's GUI guidelines, more specific a single set of guidelines. A set of guidelines which are usable anywhere not just on a single desktop, best if usable cross-platform, something like wyoGuide (http://wyoguide.sf.net/). IMO the FSG should not only standardize the binary interface but also delve into standardizing the GUI interface. This would IMO give the FSG much more importance.
People talk about how great Aqua is, but it's not exactly taking the world by storm. OpenStep is GPLd, and is really failing miserably to get a foothold against GTK, QT and Motif toolkits, as well as native X11 (QT and Lestif are GPL'd, but they have closed-source commercial counterparts for those willing to pay).
Yet a framework needs a license as wxWidgets so it can be used everywhere. Anything else is IMO ridiculous.
I could go on, and on, but I think that's enough to make my point. If you want a standard, you can't use the GPL.
That's not entirely correct, it quite much depends on what kind of object is licensed. While it may make sense to license an OS or an application GPL, anything which don't need to be linked, it's no questions a framework or a file/music format can't have the same license. In a such case the wxWindows or a BDS license is much more appropriate.
Since these boot problems are quite difficult and probably mean a no go for anybody not a though expert I really hope they were fixed before release. It probably means another delay for a few days but think it's worth.
Slashdot Mirror
I'd prefer cross-platform and/or FOSS. If I'm going to move, it will be to something that has an open format for its data files.
i d=15634627), especially wxWidgets is quite well suite for small ISVs.
[OFFTOPIC]
As you prefer cross-platform SW and develop your own software you might look at my message (http://ask.slashdot.org/comments.pl?sid=189954&c
[/OFFTOPIC]
O. Wyss
IMO financial applications are the number one sample why the Linux desktop hasn't gotten more than a few percents market share. Can you really imagine an ordinary dentist ever use GnuCash on his Windows box? Or a carpenter, or a house wife? Can you really think that such people go out, buy or install Linux on their computers? No that won't happen, not until the art of writing FOSS has changed dramatically.
. pdf).
That said financial FOSS applications will only become possible when they are true cross-platform, when they are available on Windows and MacOS as well. Yet that's not sufficient, they also have to look native and they have to feel native on any platform. Else people, who use computer as tools and not as gadgets, won't use them.
Ordinary people don't look with the eyes of a fan, the look with the eyes of an annoyed worker who wants it task done as fast as possible. None of the so far mentioned applications look acceptable in their eyes. At the current state none written in Java or with GTK will satisfy these people. The only choice which produces acceptable results are using the commercial QT or the free wxWidgets toolkit. It may sound harsh but that's the case, just listen to the complains these people bring up against FOSS applications (or read http://www.osdl.org/dtl/DTL_Survey_Report_Nov2005
Yet looking acceptable is only one step towards broad acceptance, the other step is feel acceptable. Sorry, a FOSS application following the Gnome UI guidelines does not feel acceptable on Windows, MacOSX, KDE, etc. If you port a Gnome application to another platform you have to take care of all the little details which are different, which annoy users when the don't fit. These little details are listed in the only cross-platform guidelines wyoGuide (http://wyoguide.sf.net/).
To summarize, to make a FOSS financial application successful you have to follow these simple steps:
- Design the application cross-platform, then you get enough market share.
- Write it with an acceptable look anywhere, use QT or wxWidgets.
- Care for acceptable feel anywhere, follow the cross-platform guidelines wyoGuide.
O. Wyss
You said "try wxWidgets". I did.I summarize my results:
Windows: Ok results. Application built with a wxWidgets builder.
Does this mean you built your own sample with RAD tool like wxDesigner or CodeBlocks? Then your test is more about RAD tools than wxWidgets. BTW why didn't you use wyoGuide-Demo as your starting code base?
wxWidgets on Linux is based on GTK+, therefore any wxWidgets should look equally nice as any other GTK+ application. If it doesn't you're doing something wrong. Yet I've suggested several times to wxWidgets developer to create a DirectFB (GTKDFB) port which would allow for much nicer Linux applications.
wxWidgets on Windows is based on Win32, therefore looks as nice as any MFC application and far better than any GTK+ application.
I don't know wxWidgets on Solaris (I've never tried it) but the developer capacity is rather limited. Complain to Sun if you aren't happy.
Total time preparing a "cross platform" wxWidgets application: 75 hours. But then, I *may* not know what I am doing.
Just compare once these 75 hours for 4 platforms with the one year Google needed to move GoogleEarth to Linux. Sure a full featured application may need a few hundred hours but then you're set up throughout the full development cycle.
O. Wyss
Everytime Slashdot mentions Linux, you bring up wyoGuide ...
That's not right, I'll bring wyoGuide up whenever it fits, so you won't find any mentioning in the Linux kernel discussions.
Please send feedback directly to the wyoguide-users mailing list, the chance of getting lost or forgotten is much smaller. Of course I try to not forgetting your comment here.
O. Wyss
Many people, as also I, don't know much of OSDL beside Linus Torvalds is there employed or they care for Carrier Grade Linux (whatever that means). Yet I know OSDL has done a survey about why the Linux desktop isn't a success (http://www.osdl.org/dtl/DTL_Survey_Report_Nov2005 .pdf).
t s/2005-December/000349.html).
But now what? Even if the reasons now are more than obvious does the OSDL take the next needed steps? Sure OSDL has created the Portland initiative, unfortunately these people aren't able to do anything about the most pressing matter, the first top inhibitor for the Linux desktop adoption. It might be these people simply don't know how to fix this problem albeit I've shown them one possible solution (http://lists.osdl.org/pipermail/desktop_architec
OSDL might say that "they represent the Linux community", yet OSDL isn't able to bring Linux to success, to increase its market share to a significant amount. So I would think twice if to participate in such an organization. It's sad when even the self proclaimed speaker of the Linux community can't do better.
To say it once more, without agreeing on a single set of application guidelines, guidelines which enhance the usability and the look&feel, there's no hope. All one can say is "Yet another year without a Linux desktop".
O. Wyss
I think you will find that quote is usually attributed to Albert Einstein ...
;-)
Yet Prof. Wirth used it all the time even if he wasn't original author. I wouldn't be surprised if it isn't Einstein either. So while people get known to use a quote that never implies they've invented it.
I'm using the term "schitter bis bewölkt" (sorry German) on all possible occasions and if I ever get as famous as Einstein, people would certainly attribute it to me albeit I definitely can't reclaim authorship.
O. Wyss
Why should anybody create a distributed application when a simple library API is almost always sufficient. Why making something more complex than necessary. In most cases component APIs are rather stable as soon as all the missing pieces from early beta versions are solved. Yet even if they change it's possible to handle most cases without any intermediate interface definition etc.
Prof. Wirth always said: "Make everything as simple as possible but not simpler". While Prof. Wirth as made many things too simple (to prove his statement true) any component system is yet much too complex for any locale task and many times also even for distributed tasks. I'm still waiting for a component system which is as easy usable as a simple library API.
O. Wyss
don't write code which needs several month to get accustomed! That's the same mistake as is usually done in science. This isn't limited to science or computers, it's all over the world.
Yet computer code is more sensitive to this kind of miss interpretation since neider coders nor their bosses know all the necessary "things" making their code more understandable for others. That's one important reason I still stick to the sample code within the wyoGuide guidelines (http://wyoguide.sf.net/) even if it gets rated biased towards wxWidgets. IMO it's incredible important for OpenSource as for ClosedSource to make code as readable as possible and nothing is better suited to show how than sample code.
Sure enough it would make a lot sense to have sample code for other frameworks or environments but sorry, as a spare time developer I don't have that much time.
O. Wyss
Ok, so call me an un-trendy old-fart. Let's see ..... a quick review of my resume since
1980, when I first became a programmer .....
I wrote my first Fortran program in 1974 and it work fine on an IBM computer and a few years later on a Digital PDP-11. Albeit it was my first cross-platform code, there's no way to sell it these days.
Languages were never the problem and even for the same task several different languages could be used. But you can't expect users to enter countless numbers on a console or copy back the resulting numbers. So even if you code a simple square root function you have to code a GUI around which is many times more code than the function itself. Yet if you want to sell your function depending on its kind users expects sufficiently good look&feel.
O. WyssActually my wife has worked with Linux for about a year with Linux in a dual-boot environment because her computer wasn't fast enough under Windows. But after she got a new one she switched back to Windows-only because of the following reasons:
My wife uses exclusively the following applications:
- A browser (Mozilla/SeaMonkey)
- An office (MSOffice, Excel/Word)
- A mailer (Outlook)
- A music recorder/editor (Audacity)
The reason my wife used the dual-boot Linux was, Audacity under Windows couldn't then record music with a sufficient acceptable quality. After this limitation was lifted she dropped Linux in favor of Windows because there isn't a mailer like Outlook on Linux. She didn't object using OpenOffice albeit the 1.1 version wasn't a real match to MSOffice at that time. You might wonder why my wife insisted on using Outlook. Simply at office she is pressed to use Outlook and didn't like using a second mailer.
To make Linux more attractive for my wife I don't suggest to build yet an Outlook clone on Linux albeit this would help. Yet the real solution would be building a cross-platform mailer which outperforms Outlook on Windows. Outperforms not in the eyes of a Linux-Fan but in the eyes of an ordinary business worker doing his job.
O. Wyss
Read Slashdot of course!
Yes. Slashdot has quite a reputation for attracting knowledgeable people, yet be aware that some are rather biased towards OpenSource. And don't forget that people who do OpenSource (me including) have a rather absolute opinion. So as long as you are a little sceptic you should be able extract the trends.
As you mention Java you may well notice that currently any Java discussion always tends toward flame wars. Flame wars are always signs that something isn't good, that the there isn't an uphill trend. Flame wars always arises when the future (a trend) isn't going as wished.
I'm probably much biased but IMO the future trend in software development is "cross-platform". So far for many years you could do resonable cross-platform development only with Java. Today you can equally well do cross-plaform development with AJAX or with wyoGuide (binary applications, http://wyoguide.sf.net/). So regardless which of the different technology takes the lead, cross-platform development will increase to the point where single-platform development won't be accepted.
O. Wyss
Take away Access and they'll do it all in Excel
Yep I've seen this too. It's never a good thing to just take something away without giving an alternative.
Beside as soon as people starts involving the IT, it's inportant enough so they'll pay for it. If not, it's not worth the effort. We have made very good experience with applying a price tag to such kind of stuff.
O. Wyss
Just in case ISVs want to develop efficiently cross-platform application there's a helpful article at NewsForge (http://software.newsforge.com/software/05/04/22/1 859243.shtml?tid=89).
O. Wyss
While for developers code may count it doesn't for users. If native code dies is only a questions if developers are able to build easily native code as if none-native code. As long as there are useful tools around. An these days this can be safely answered with yes.
On the other side users care at lot for native GUIs. As long as users are able to distinct e.g. a Java application from a binary (native code) application within 5 min. of using, vendors selling SW have to care for native code which produces far better native GUIs than none-native code.
Yet the SW industry does not work as in other branches. Since development costs is the driving force other considerations have to be taken into account. So for inhouse development, where users complains don't matter that much, none-native Java is favored and forced. On the other side in the free world and the commercial vendors there's no alternative to native GUIs therefore native code.
Yet the future of native versus not native isn't decided so far. With the hyped AJAX there's a none-native technique on the rise while with wxWidgets/wyoGuide (http://wyoguide.sf.net/) there's a now superb cross-platform native technique available. So it might be that in the end there's a draw between native and none-native going along very well.
O. Wyss
I'm amazed, even the title of this story is wrong. For browsing any tool who can display code is sufficient. Sure tools which can syntax highlight or have name reference lookup are better but the real issue is to find the source code first hand.
Whenever I search for a solution I first go to http://koders.com/ but their index seems a little limited. Still just try once looking up "wxSingleInstanceChecker". There are others like Koders.com but I've forgotten their names. Next I try to think of a most fitting statement for code and feed it to Google (e.g. "class App: public wxApp", but beware of the white space) which at least returns some hints which project might have source code available. Yet the free text search isn't very suited for searching code since it produces too much wrong results.
When I've found a project which might have fitting code I either look into their LXR if it's available or simply download the source tarbal and use a decent editor (e.g. http://freshmeat.net/projects/wyoeditor/).
It's said that none of the current CVS web tools are searchable, nor that Google is able to restrict results to CVS pages, else it would be much easier to search for source code.
O. Wyss
"Wouldn't it be great if everything worked in a way which is most convenient for me"
Very accurate translation, thank you very much.
O. Wyss
You could aw wella ask why it matters if Windows is in non-free, in both cases it's about law, not ideology.
Three weeks argumenting back and fore isn't law or at least not understand by the posters.
O. Wyss
What I meant, was exactly which aspect of the Linux marketplace is Debian "leading" in ?
You seem to be not a developer else you would know what "leading" I meant. For many years OpenSource developers preferred Debian over all other distributions because of various reasons. IMO the most important reason is, you could easily mix a stable distribution with experimental packages which sometimes were essential for development. Yet Debian has fallen back in providing enough recent packages and yet Ubuntu can equally be intermixed. So Debian is loosing upstream developers. I might overestimate or you might underestimate what this effect means for Debian but it's definitely not good.
O. Wyss
It's sad that Debian always provoke such ideologic wars as if no more important work and decisions are needed. What does it matter if Java is in none-free or else. Who cares much except users who just want to download a distribution as easy as possible.
IMO there are bigger issues at stake for Debian. The upcoming Ubuntu distribution is threatening to replace Debian as the leading distribution. The GPL v3 does nowhere fit in well into the DFSG. The Linux desktop while each year announced does not take place. It seems Debian people just fights these wars so nobody notices the real threads Debian is faced with.
I've myself switched to Ubuntu since I'm not happy with Debian anymore and I'm most probably don't write about the Linux desktop in the Debian mailing lists anymore etc. Any I'm quite sure quite a lot of others do the same. I've simply got tired of all this ideologic stuff.
O. Wyss
Unfortunately for the editors, several thousand someones just submitted Firefox stories in the span of about fifteen minutes.
Nevertheless the editors should care more for qualitiy than quantity but I agree they have to keep an eye on the numbers as well.
O. Wyss
I would argue not using VB on the basis that it is not cross platform. ...
I second that, if you become successful with your product sooner or later you'll be confronted with this question. As soon as you want to sell your product to governments or city administrations you'll be faced with this question.
Yet VB definitely has no long time future and you have to switch anytime. If you surely know you always stay in the Windows environment you might choose C# but I wouldn't advice that. You can equally well choose to switch to C++/MFC (as somebody else advised) and still have the road open to anything. From C++/MFC it's quite easy to switch to C++/wxWidgets which you should target in any case in the long run. With wxWidgets your code almost instantly becomes cross-platform without much extra work. All you have to do is keep an eye on wyoGuide (http://wyoguide.sf.net/), it advises you how to code so your product becomes cross-platform even if you never actually try it out.
So before you switch design your development strategy first. Since development always needs time, sometimes years, and you have to support your product even longer. So IMO moving slowly towards C++/wxWidgets pays off in any case. Any good developer knows C++ and wxWidgets is quite a full featured framework.
There are more ways to develop good products but I simply can't advise them. Java with all its addins isn't easier to code than C++ anymore, faster development time has become a joke and Java applications are easily spotted by the users and definitely not liked. The QT framework is another alternative but there seems to be no free cross-platform applications and when you look how long it takes to port GoogleEarth to Linux I'm loosing fait. Yet the GTK+ framework, forgive me if I say it so blunt, might be useful for free SW but in no case for commercial development.
O. Wyss
Have you posted a bug about it in Launchpad?
Read the Ubuntu forum message and you will know.
O. Wyss
It might be that the LSB makes life easier for distribution but does it also have an effect for developers and users? I don't remember I ever have looked into the LSB when designing and coding an application nor when distributing source files. And I'm quite sure most users don't even know that the LSB exist. While the LSB is very important for the binary distribution, it's influence on a Linux system is rather limited. Yet the FSG only cares for the LSB and therefore it's importance is also rather limited.
A useful Linux system needs some user's GUI guidelines, more specific a single set of guidelines. A set of guidelines which are usable anywhere not just on a single desktop, best if usable cross-platform, something like wyoGuide (http://wyoguide.sf.net/). IMO the FSG should not only standardize the binary interface but also delve into standardizing the GUI interface. This would IMO give the FSG much more importance.
O. Wyss
People talk about how great Aqua is, but it's not exactly taking the world by storm. OpenStep is GPLd, and is really failing miserably to get a foothold against GTK, QT and Motif toolkits, as well as native X11 (QT and Lestif are GPL'd, but they have closed-source commercial counterparts for those willing to pay).
Yet a framework needs a license as wxWidgets so it can be used everywhere. Anything else is IMO ridiculous.
I could go on, and on, but I think that's enough to make my point. If you want a standard, you can't use the GPL.
That's not entirely correct, it quite much depends on what kind of object is licensed. While it may make sense to license an OS or an application GPL, anything which don't need to be linked, it's no questions a framework or a file/music format can't have the same license. In a such case the wxWindows or a BDS license is much more appropriate.
O. Wyss
I happen to test around several boot problems the last few weeks I've summarized just here
3 67#post1062367
http://www.ubuntuforums.org/showthread.php?p=1062
Since these boot problems are quite difficult and probably mean a no go for anybody not a though expert I really hope they were fixed before release. It probably means another delay for a few days but think it's worth.
O. Wyss