“The good news for Linux as an operating system for the desktop–as opposed to the server–is that it is set to become No. 2 after Windows in the next year or so. The bad news is that its growth does not look to be as explosive as some advocates might have hoped.” Read more at ZDNews.
I had been hoping to hear somthing like that soon. This is indeed good news for the Linux community. Greater adoption means greater support from hardware and software makers. Already, 2002 has seen an unprecidented level of support for Linux in products of all kinds.
Anyone think Linux will beat out the Mac for the #2 spot on the desktop? For me personally, I think the Mac would be a better choice for me on the desktop than Linux (because of all the audio apps I use), but I don’t really want to spend the money on Apple hardware to esentially do the same thing I’m already doing in Windows, so I’m willing to wait for Linux to catch up, and I think a lot of people feel the same way I do.
I recall that the last HP printer I got contained a full color ‘Quick Start’ guide for PC & Mac. If Linux jumps ahead of the Mac, then my next printer may just have such a poster for Linux too (and accompanying drivers if needed). If that happens, it would be a HUGE victory for desktop Linux, combined with the inroads developers are making on usability.
The only thing it would need at that point is to fill in many of the gaps currently present due to the lack of certain ‘speciality’ apps, and you’ve got a real contender.
I had hoped for greater share and faster growth. But as the article says growth is hinderd by lack of apps. Good browsers, mail clients and to some degree office suites are not going to cut it alone. Also good apps in other areas such as gaming, photo, multimedia and file sharing (P2P) are needed.
Sun wants to get the shovel under MS. Then why doesn’t Sun develop all of the above apps in a coherent suite of Java apps? J2EE on Linux is already pretty well established in the back room.
Then why doesn’t Sun develop all of the above apps in a coherent suite of Java apps?
Well, I agreed with pretty much everything you said until you got to this part. Java??? There are many uses for Java … GUI apps on the desktop are not one of them, at least not yet.
I have ran a few Java desktop apps and, although they’re not as slow as say … HotJava 1.0, they don’t seem nearly as fast or responsive as native bretheren either.
I agree with you on that Darius. LimeWire isa good example. I covers the p2p issue quite well….but, because it runs on java . Although it runs great it never seems as responsive as GTKgnutella.
While GTKgnutella also runs very well it lacks the polish and refinment of LimeWire.
Refined applications are going to become an issue as people become spoiled by the likes of OpenOffice, Mozilla, gnome(itself as a desktop) and KDE. And newer users are accustumed to refined looking programs (reguardless of the underlying quality) Now Gimp is getting there but slower where functionality is greater concern than polish.
To people like myself polish and refinment is not an issue. I understand the diference between a release and a alpha, beta, and release candidate. However to joe blow Aoler that polish and package makes the sale.
hopefully the numbers will increase and start atracting the likes of macromedia, adobe ect. which comes to another issue that is getting better. Linux users have to be willing to buy software from companies that provide software they want. good example:
Unreal tourn. 2003
I paid for this for the sole reason that it rocks (great quality) and runs on Linux.
Distros is another, I buy them and get the added benifit of buying opensource.
Linux at its current stage is too fregmented – a nightmare for tech support – except for a few BIG players backed with huge amount of resource.
To change the situation, either Linux can find a way for unified software deployment/configuration, or there comes a monopoly player that can force such a way into every one’s throat – smells like redhat, huh ?
I agree. Java is slower than compiled code and always will be. I use jEdit for C/C++ and am satisfied with its speed performance. But, the fact that MS also is betting on a similar slower techology (.Net) whereby a program is compiled for a VM and not a real machine should give SUN some confidence to move ahead. Hardware will become faster to the point where it doesn’t matter. Fast is fast enough. Once .Net is on the desktop then it is too late for SUN. (HotJava is at 3.0 but about 5 years old).
So Java with its Write Once, Run Everywhere burden is unlikely to beat .NET on M$’ turf.
It is growing in a remarkable pace. You must remember that while the % may be small, it represents a large amount of users.
“To change the situation, either Linux can find a way for unified software deployment/configuration, or there comes a monopoly player that can force such a way into every one’s throat – smells like redhat, huh ?”
I also think that Linux is hindered at the current time by the fact that there is no UI standard like there is for Windows. There are also no decent GUI toolkits that can be had for commercial software without spending a huge amount of money. QT costs too much (like $2,000 per license for the Enterprise version), and GTK is ugly and a pain in the ass to program with.
Linux really needs a standard GUI toolkit that looks decent, is relatively easy to work with, and doesn’t require you to mortgage your house to buy it.
As far as Linux beating out Mac as the #2 desktop, I doubt it. Steve Jobs will probably pull something out of his hat before that happens, which I still think could possibly be an x86 port of OS X.
Actually, Microsoft is betting on .NET to kill Java :-). Not much hope for Sun is there? Well, on a more serious note, there is actually some use to languages like Java and C#. However, Microsoft is not positioning them to replace compiled apps – remember, they are also investing a lot in Avalon.
“I had been hoping to hear somthing like that soon. This is indeed good news for the Linux community.
Greater adoption means greater support from hardware and software makers. Already, 2002 has seen an
unprecidented level of support for Linux in products of all kinds.”
I think a commercial software company will want to know, not the
number of users, but the software sales in dollars on each platform.
Linux could have several times as many users as Mac, but they may well
not buy much software, especially expensive programs such as
Macromedia or Adobe or Newtek sell.
Many Windows software companies consider the Mac market too small to
bother with.
Well since it has been proven fairly easy to recompile source for Macintosh computers running versions of X on top of Aqua, if there is an increase in good OSS Linux Apps, Mac users might see a lot more on their screen.
There’s always another angle you have to look at
.
India and China are both planning to make Linux their “national” operating system. The Chinese have developed “Red Flag Linux” and the Indians, despite Bill Gates’ trip there to bribe them with educational grants, are in love with Linux as well.
These two countries alone are several million user currently, guaranteed to grow to… a billion or more users, as soon as every household has a computer. Windows is dead in Asia.
Taiwan and Korea are developing all sorts of Palm lookalike PDAs that run embedded Linux.
Parts of Europe, such as some German and Scandinavian government agencies, are standardizing on Linux as well. The writing’s on the wall for Microsoft.
“Gtk 1.x was indeed very ugly, but it sounds like you haven’t tried 2.x yet.. Just _try_ using it for _more_ than 2 weeks and you’ll find that it’s actually pretty good.”
I have tried it. Granted the widgets look better. But its still a pain in the ass to program with. It’s not even object oriented (for those who don’t know, GTK is a C toolkit and not a C++ one). I suppose there is nothing inherently wrong with a GUI toolkit that is not object oriented, but not having window classes and such makes for a not so easy transition for people who are used to working with almost any other GUI toolkit. Yes, I know there are C++ bindings for GTK. I haven’t messed with them yet. Sometime when I get some free time, I will have to try it.
My toolkit of choice most of the time these days though is wxWindows because I do a lot of Python programming. wxPython is great and exists for X Windows, Windows, and Mac OS X, so producing cross platform programs is a breeze. Of course, if I later want to rewrite in C++, most of the UI code will be the same for wxWindows as for wxPython.
Unfortunately, wxWindows doesn’t seem to be that commonly used in major projects. I’ve been toying with the idea of writing a desktop environment in wxWindows. But I don’t have the time to work on it right now.
I also like the FOX toolkit, but its not as well documented.
[Tongue in cheek]
Paying for opensource products?! No, free MUST mean free as in beer. If they aren’t going to spend hundreds of thousands of dollars working on it to just give it away, it’s evil and they’re anti opensouce and don’t give back to the community!
Whine, snivvel, piss, bitch, moan
[end Tongue in Cheek]
I do the same thing actually. I’ve bought and paid for software that I can just download in minutes over my cable modem, much to the horror and chagrin of the nobody should ever pay anything for opensource crowd.
“You PAID for a copy of Star Office 5.2? But you can just download it and compile it for free.”
“Call me a freak, but I definatly want to throw a few bones to a company producing a reasonably priced office suite and manual that liberate me from buying anything new from THAT company. Who knows, it could make the difference between them puting out a newer, better version next year, and letting it die an untimely death.”
>> “Yes, I know there are C++ bindings for GTK. I haven’t messed with them yet. Sometime when I get some free time, I will have to try it.”
I should have mentioned that I’m only using plain Gtk2 when writing for my theme engine (not released yet), but it’s still not that worse.. (compared to Gtk1).
I’ve been using the C++ bindings (Gtkmm2) since one month now and it’s much more like other toolkits this way (with a nice twist).
I think it’s actually easier than other tk’s using Gtkmm2.
(the optional STL-like Helper-API for pushing/popping menu items (example) in a Gtk::Menu is also very nice)
Also, there’s no need to ref/unref things yourself.
Smart pointers all the way!
>> “My toolkit of choice most of the time these days though is wxWindows because I do a lot of Python programming. wxPython is great and exists for X Windows, Windows, and Mac OS X, so producing cross platform programs is a breeze. Of course, if I later want to rewrite in C++, most of the UI code will be the same for wxWindows as for wxPython.”
I was using wxWindows (C++, no Python or anything like that) before I decided to use Gtkmm2 instead but for some reason wxWindows didn’t feel right to me.
There wasn’t much consistency between the Win32 and Gtk(2) ports, scrolling is done badly (not per pixel, for some widgets), it isn’t able to handle an alpha channel in images (very important for what I’m doing), many widgets in wxWindows flicker like ****, very bad font support, some widgets are implemented using custom code and look pretty out of place .. and the list goes on and on. :/
This all could be fixed of course, but that’s not my job unfortunately.
(committing bugfixes is fine with me, but replacing complete font managing systems… (for example))
I’ve used MFC, plain Win32, Qt, wxWindows and some other toolkits and I think I’m going to stay with Gtkmm
It does the job, and it does it very well.
Once this is released I think it is likely to be the obstacle to immediate success. There’ll be a fracture of those using the 2.4 kernel and those using the 2.6 kernel and from what I’ve read the API has changed quite a bit.
Having users split across the two kernels is likely to cause a lull in developement as people alter old code to run on the new kernel.
Once that little bridging period is over and everyone is running on the same API I think Linux is going to move forward quite a bit this year. Just wish someone would write a decent XFree86 replacement.
>> “Just wish someone would write a decent XFree86 replacement.”
No, it should be improved, not replaced (which would take years.. look at DirectFB or Berlin for example)
>> “No, it should be improved, not replaced ”
Nope, it needs replacing by something that’s designed to be a desktop and not a network transparent gui utility.
Gtk, though implemented in C, is object oriented. Also, there are C++ bindings available. Another toolkit that is both object oriented and cross-platform is the Fox toolkit.
Another toolkit that is both object oriented and cross-platform is the Fox toolkit.
I’ve worked with the Fox toolkit, and I like it a lot. It’s main problem right now is that its documentation needs work.
“I’ve used MFC, plain Win32, Qt, wxWindows and some other toolkits and I think I’m going to stay with Gtkmm
”
I’d like to give Gtkmm some serious consideration. The only problem I have is that unfortunately, most of the programming I have to do is in Windows. And Gtk is still rather flakey under Windows. Also, I don’t think there is a Python binding available for Windows.
I take that back. There are some PyGtk bindings for Windows. Suppose I will download them and play with them for awhile.
I’ve worked with the Fox toolkit, and I like it a lot. It’s main problem right now is that its documentation needs work.
You can say that again. I took 2 days to figure both Qt and Python out, wrote a small app with the help of Qt’s documentation, and a Python tutorial on the web. I wish I didn’t delete that app….
I would just like to point out that Red Flag isn’t the only distribution in mainland China. However, it is the biggest company, but a private company, not a state-owned one.
There are no technical reasons to not use linux on some desktops. I only use linux to browsing and email, for writing CDs, for CD ripping and mp3 generation, for gaming (for some games like Quakes, Unrela Tournament, Castle of Wolfenstein, etc), for C/C++ programming, for LaTeX editing (it is easier to make a latex environment on linux than on Windows) and even run some WIN32 applications like Kazaa Lite and M$ Office 97 using wine.
The only problem is political. The american software companies and Microsoft are allieds and they don’t want competition. With linux being adopted in the third world and Europe, they will have fair competition. American software is also very expensive for third world countries because their price is based on US dollars. In my country (Brazil), a single copy of M$ Office Professional is 5 X minimum mensal salary and a copy of Windows XP is 3X. This is the reason to rapid spreading of linux in Brazil. We can have national linux distributions (like Conectiva Linux) and native development of linux applications.
Yup…for example thizlinux which is made in Hong Kong is also another distro aviable to those living in China. I am pretty sure there are more then just these two distros. floating around in China.
http://www.chaintech.thizlinux.com/
I agree with the author that ELX is a vastly under-rated distro.
Also, Lycoris does have its IRIS download center now, although it doesn’t have a whole lot of applications yet.
>> “I’d like to give Gtkmm some serious consideration. The only problem I have is that unfortunately, most of the programming I have to do is in Windows. And Gtk is still rather flakey under Windows. Also, I don’t think there is a Python binding available for Windows.”
Gtk 2.2 should have improved support for Windows and it’s being considered to make it an official port.
The only problem is that Gtk(2) looks really out of place on Windows, but that’s easely solved using a theme engine.
“Gtk 2.2 should have improved support for Windows and it’s being considered to make it an official port”
Hopefully. That would be nice to have a good alternative to MFC that doesn’t cost a fortune to buy.
The main problem with Gtk on Windows that hopefully they will fix soon is that multithreading isn’t safe. I’m pretty sure they only recently fixed that in the X version. I know that was a fairly common complaint before Gtk 2.
“The only problem is that Gtk(2) looks really out of place on Windows, but that’s easely solved using a theme engine.”
It looks out of place now, but maybe that will change. If Gtk can keep improving and Linux can gain enough desktop marketshare that desktop app vendors start to take it seriously, Gtk might become the toolkit of choice for vendors that want to produce their apps for multiple platforms. Maybe Gtk will even become the toolkit of choice for Windows programming over Microsoft’s libraries. It’s got a long way to of course, before that though.
There have been some major improvements made in Windows support though. For example, the latest Gtk Windows ports can set their colors based on the current Windows color scheme being used. That in itself is a major step forward that makes it so it doesn’t look quite so out of place.