The tides of the computer world may be turning. The part of this article that I thought was most interesting was actually the link to the story about the pending announcement of Next Generation Windows Services (NGWS). It seems that MS will be shifting strategies towards running applications off of servers. Serving up applications like so many web pages. Charging for them by subscription. A new revenue model. Or should I say a new monopoy model!
Just when we are on the verge of having a very high quality free desktop operating system and office package combination (in the form of KDE/GNOME, and all the various office suites). Coincidence? Maybe MS has given up the idea of controlling the desktop. If their losing it is a forgone conclusion, why not focus on taking over the server market.
Unix still has a good share of the server market, but I can see this changing if we are not careful. How? By introducing products like their "BizTalk 2000", they effectively create defacto standards for e-commerce. Sure, it's all XML driven. Sure Linux apps could read it and maybe even talk to it. But so what? There are no linux apps that do it because MS won't write 'em for Linux. That's the point. If I'm running a business, I want to be compatible with the other businesses. I want to be able to talk to them and do business. So I'm going to run what they run. If MS leverages their current server market share to covert it to e-business server share then they will create the defacto standard. Linux representitives will have virtually no say in this new area.
It seems Microsoft has been busy! And I thought all their time and money was being wasted fighting the DOJ.
How about BitKeeper. I've never used it, but it looks pretty nice. I've been meaning to try it out. It's been in beta for a while. It has lots of GUI tools and support for 'Lines of Development' (LOD). It aims to be the version control system for the Linux kernel. A lofty goal indeed.
The licensing is a bit wacky, as it's not exactly free.
My advice to you, friend, is to use an account, go to the preferences page, find the "Exclude Stories from homepage" section and check "Mozilla". Because you obviously don't care. Thanks. But personally I think you'll miss out on information about what is becoming one really cool application.
> This quote seems to sum it up: > I send a letter to a friend over postal mail. > In the mail, i say "I have a package for you. > I'll send it via fedex". I > then send him a kiddie porn video. Is the postal > office criminally negligent for allowing the > transmission of child > pornography?
Sorry, that doesn't quite cover it.
Because what's really happening is your are writing to the postmaster general and saying: "who has kiddie porn?". The post master general writes back and says: "Bob has kiddie porn. Here's his address. By the way, getting kiddie porn is illegal." Then you write to Bob and say: "Bob, will you send me a free copy of your kiddie porn?"
Good points. I don't think there really is much relationship between SE and Open Source.
As an individual, I could use some SE methods to develop an Open Source app, but who does that? No one I know of. I actually think that it might work and produce some good quality stuff if they did. Especially if all the people working on it have access to the same tools.
One important issue is the tools. KDevelop may be nice, but how does it help you do software engineering? It doesn't. What we need is quality CASE tools. ArgoUML shows promise but is lacking a lot of features and is still early on in its development. Together/J is pretty nice, but you need the commercial version to get all the features.
I have never seen an open source app that deals with requirements and design documentation. The thing we like to be able to do in industry is trace (follow) requirements to their implementation in design and code.
And then there's version control. CVS doesn't have all the features of something like ClearCase, but bitkeeper is looking good.
Why does so much commercial software suck when compared to open source? Probably lots of reasons. Market timing pressures, poorly documented legacy code, bad management. But maybe the most important one is lack of large scale peer review. In open source there are a lot of fresh faces who are free to come in and say:
"This design sucks!! We should scrap it and start over and we will come up with something better. Fine, i'll do it myself in a few hours."
You'll almost never see something like that said (out loud) in industry.
Maybe you had a bad professors for your courses. Mine really knew his stuff. In fact, he recently started a Software Engineering major at MSOE.
There were a few good things about my SE class. (I only had one). One thing was that we were split into teams (not of our choosing). This simulates the real world. We also had project to work on where the professor was the client. And we did the project in 2 phases. (We only had 10 weeks!) The 2nd phase added some requirements and we got to see how good our design from the first phase was. We got to use Rational Rose, which is a pretty nice CASE tool. Once you put your model in as UML, you click the "Generate" button, and Bam! Out comes the skeleton code in either C++ or Java. We used C++ and put a MFC based front end on our app. Doing the actual programming was the easy part!
Then as an outside project I took the back end code and wrote a KDE front end for it!. And I did it in you a couple of hours!! I think this shows the importance of creating a design that is as non interface-specific as possible. (Depending on the purpose of the app, of course). One thing I like to do is think of what the back end code would have to look like so that the interface could be either text mode or GUI. It makes you think! Perhaps a good example of this is Linuxconf, with a text, X, and web interface, but I haven't looked at the code at all.
The SE course I took was probably on of the hardest (most work) courses I had in my four years, but it was also the most interesting.
I do agree that SE is a young field. Compare it to something like Architechtural Engineering. Architechture has been around for thousands of years. Computers have only been around for, what, about 50. And the idea of actually doing software engineering is even younger.
Hopefully the good ideas (I hate the phrase "best practices") in Software Engineering will stick and we will have better software to show for it soon. I agree with another poster though in that market pressures make this difficult for commercial software development.
There used to be a slashdot page where we could see the daily hit count by browser and OS. While sometimes depressing (2/3's browsing from Windows!), it was very interesting. Is there any chance we will see this again? Is this now information that you feel you need to keep private for some reason? What about the number of registered slashdot users? Could we find that out?
This is one of the great things about Linux in general. Every distribution is free to bring in new features and tweaks to appeal to the users. And if they are good ideas, they will catch on.
I think a selectable security level is a great idea. The trade off is decreased usability for increased security. I can see tons of uses for this setting. For example, setting which services are running. (The default Redhat installations have a lot of things running and it is really scary.) Also setting whether or not to create a non root user account, and setting who can mount drives and access devices.
A really secure installation might not even have telnet, and instead go with ssh. It could have tripwire setup for use too! And it could have tcp/ip syn cookies on. It could log all odd port connections (somehow?) The posibilities are endless! I know how paranoid I am. And I'm not the only one!
The only possible problem I see with it is that it adds another variable to examine with bug reports. Certain problems may only happen with certain security levels, and this adds some complexity to things. But maybe I don't know what I'm talking about and it will all be manageable. (I hope so!)
I'm not sure why this is. Maybe they just didn't get around to implementing it yet. It seems like it would be a simple (and useful) thing to add a check box in the preferences, but I didn't see one.
Same here, pretty much. I learned C using the DOS based Borland C/C++ 3.1.
One very nice feature that it had (if I recall correctly) was smart syntax highlighting. For example: If I have a #define and I use it later, I want to be able to see by its color that it is a #define, not just another variable. Another example would be having all function names (regardless of location) be the same color. Does anyone know of any free/open source editors that can do something like that? Everything I have seen so far just seems to be based on glorified regular expressions. -John
Check out the penguin on the about page. I like it and I think Linux needs a 3D mascot. I also like the whole site overall. Nice work!
-John
License Shmicense. KDE and GNOME both suck
on
KDE 1.1 is out
·
· Score: 1
Well, I want a coherent desktop environment, but not at the price of system stability and usability. I guess I want to have my cake and eat it too. Is that too much to ask?!
I have not tried any recent GNOME offerings, but I did install KDE 1.1pre2. I only have 32MB of RAM and 18MB of swap. Yes, I know, my fault for setting it up wrong so long ago. Linux needs something like HPUX's logical volume manager so I can just add to the swap.
Anyway, I run KDE, start up Netscape, Gimp and two of those Kvt terminals. Those stupid terminals take up 4MB of RAM. What the hell?! Well, as you might guess, my system starts swapping like there is no tomorrow. It becomes completely useless and I have to hit the reset button:(. Well, after that experience it was back to FVWM for me! I appreciate a window manager that starts up in about 5 seconds;)
Slashdot Mirror
Just when we are on the verge of having a very high quality free desktop operating system and office package combination (in the form of KDE/GNOME, and all the various office suites). Coincidence? Maybe MS has given up the idea of controlling the desktop. If their losing it is a forgone conclusion, why not focus on taking over the server market.
Unix still has a good share of the server market, but I can see this changing if we are not careful. How? By introducing products like their "BizTalk 2000", they effectively create defacto standards for e-commerce. Sure, it's all XML driven. Sure Linux apps could read it and maybe even talk to it. But so what? There are no linux apps that do it because MS won't write 'em for Linux. That's the point. If I'm running a business, I want to be compatible with the other businesses. I want to be able to talk to them and do business. So I'm going to run what they run. If MS leverages their current server market share to covert it to e-business server share then they will create the defacto standard. Linux representitives will have virtually no say in this new area.
It seems Microsoft has been busy! And I thought all their time and money was being wasted fighting the DOJ.
The licensing is a bit wacky, as it's not exactly free.
My advice to you, friend, is to use an account, go to the preferences page, find the "Exclude Stories from homepage" section and check "Mozilla". Because you obviously don't care. Thanks. But personally I think you'll miss out on information about what is becoming one really cool application.
> This quote seems to sum it up:
> I send a letter to a friend over postal mail.
> In the mail, i say "I have a package for you.
> I'll send it via fedex". I
> then send him a kiddie porn video. Is the postal
> office criminally negligent for allowing the
> transmission of child
> pornography?
Sorry, that doesn't quite cover it.
Because what's really happening is your are writing to the postmaster general and saying: "who has kiddie porn?". The post master general writes back and says: "Bob has kiddie porn. Here's his address. By the way, getting kiddie porn is illegal." Then you write to Bob and say: "Bob, will you send me a free copy of your kiddie porn?"
As an individual, I could use some SE methods to develop an Open Source app, but who does that? No one I know of. I actually think that it might work and produce some good quality stuff if they did. Especially if all the people working on it have access to the same tools.
One important issue is the tools. KDevelop may be nice, but how does it help you do software engineering? It doesn't. What we need is quality CASE tools. ArgoUML shows promise but is lacking a lot of features and is still early on in its development. Together/J is pretty nice, but you need the commercial version to get all the features.
I have never seen an open source app that deals with requirements and design documentation. The thing we like to be able to do in industry is trace (follow) requirements to their implementation in design and code.
And then there's version control. CVS doesn't have all the features of something like ClearCase, but bitkeeper is looking good.
Why does so much commercial software suck when compared to open source? Probably lots of reasons. Market timing pressures, poorly documented legacy code, bad management. But maybe the most important one is lack of large scale peer review. In open source there are a lot of fresh faces who are free to come in and say:
You'll almost never see something like that said (out loud) in industry.Maybe you had a bad professors for your courses. Mine really knew his stuff. In fact, he recently started a Software Engineering major at MSOE.
There were a few good things about my SE class. (I only had one). One thing was that we were split into teams (not of our choosing). This simulates the real world. We also had project to work on where the professor was the client. And we did the project in 2 phases. (We only had 10 weeks!) The 2nd phase added some requirements and we got to see how good our design from the first phase was. We got to use Rational Rose, which is a pretty nice CASE tool. Once you put your model in as UML, you click the "Generate" button, and Bam! Out comes the skeleton code in either C++ or Java. We used C++ and put a MFC based front end on our app. Doing the actual programming was the easy part!
Then as an outside project I took the back end code and wrote a KDE front end for it!. And I did it in you a couple of hours!! I think this shows the importance of creating a design that is as non interface-specific as possible. (Depending on the purpose of the app, of course). One thing I like to do is think of what the back end code would have to look like so that the interface could be either text mode or GUI. It makes you think! Perhaps a good example of this is Linuxconf, with a text, X, and web interface, but I haven't looked at the code at all.
The SE course I took was probably on of the hardest (most work) courses I had in my four years, but it was also the most interesting.
I do agree that SE is a young field. Compare it to something like Architechtural Engineering. Architechture has been around for thousands of years. Computers have only been around for, what, about 50. And the idea of actually doing software engineering is even younger.
Hopefully the good ideas (I hate the phrase "best practices") in Software Engineering will stick and we will have better software to show for it soon. I agree with another poster though in that market pressures make this difficult for commercial software development.
There used to be a slashdot page where we could see the daily hit count by browser and OS. While sometimes depressing (2/3's browsing from Windows!), it was very interesting. Is there any chance we will see this again? Is this now information that you feel you need to keep private for some reason? What about the number of registered slashdot users? Could we find that out?
This is one of the great things about Linux in general. Every distribution is free to bring in new features and tweaks to appeal to the users. And if they are good ideas, they will catch on.
I think a selectable security level is a great idea. The trade off is decreased usability for increased security. I can see tons of uses for this setting. For example, setting which services are running. (The default Redhat installations have a lot of things running and it is really scary.) Also setting whether or not to create a non root user account, and setting who can mount drives and access devices.
A really secure installation might not even have telnet, and instead go with ssh. It could have tripwire setup for use too! And it could have tcp/ip syn cookies on. It could log all odd port connections (somehow?) The posibilities are endless! I know how paranoid I am. And I'm not the only one!
The only possible problem I see with it is that it adds another variable to examine with bug reports. Certain problems may only happen with certain security levels, and this adds some complexity to things. But maybe I don't know what I'm talking about and it will all be manageable. (I hope so!)
Solaris went from version 2.6 to version 7!
:)
It's basically just a form of marketing. Gotta keep up with the other guys, or try to exceed them for a little while.
Given this news, I bet Redhat will go to version 7 too.
I'm just happy to have a new cutting edge distribution about every 3 months
Funny, I just looked this info up.
Sun Ultra 10
------------
proc: UltraSPARC IIi
speed: 440MHz
ram: 256MB
disk: 9.1G EIDE 7200RPM
graphics: PGX24 (24-bit)
OS: Solaris 7 (2.7)
sound: none?
monitor: 17in (Sun monitor _required_)
network: 10/100
cdrom: 32x
cost: $6350
source: www.sun.com
VA LiNUX StartX MP
------------------
proc: Two Intel Pentium IIIs
speed: 500MHz (each)
ram: 256MB
disk: 18.2GB SCSI
graphics: Matrox G200 AGP
os: VA Linux OS 6.0
sound: Soundblaster 16
monitor: (extra, adds ~$400)
network: Intel EEPRO 10/100
cdrom: 40x
cost: $3443
source: www.valinux.com
The same, better, or twice the system for almost half the money?!?!?
What am I missing?
I don't see how Sun gets away with this. Maybe the won't for much longer?
(PS: Slashdot need the pre tag in html mode!)
I'm not sure why this is. Maybe they just didn't get around to implementing it yet. It seems like it would be a simple (and useful) thing to add a check box in the preferences, but I didn't see one.
There is an online validator at:
http://www.stg.brown.edu/se rvice/xmlvalid/xmlvalid.shtml
One very nice feature that it had (if I recall correctly) was smart syntax highlighting. For example: If I have a #define and I use it later, I want to be able to see by its color that it is a #define, not just another variable. Another example would be having all function names (regardless of location) be the same color. Does anyone know of any free/open source editors that can do something like that? Everything I have seen so far just seems to be based on glorified regular expressions. -John
-John
I have not tried any recent GNOME offerings, but I did install KDE 1.1pre2. I only have 32MB of RAM and 18MB of swap. Yes, I know, my fault for setting it up wrong so long ago. Linux needs something like HPUX's logical volume manager so I can just add to the swap.
Anyway, I run KDE, start up Netscape, Gimp and two of those Kvt terminals. Those stupid terminals take up 4MB of RAM. What the hell?! Well, as you might guess, my system starts swapping like there is no tomorrow. It becomes completely useless and I have to hit the reset button :(. Well, after that experience it was back to FVWM for me! I appreciate a window manager that starts up in about 5 seconds ;)