For me, the biggest threat to Linux is the hardware/driver support.
I gave up on Linux for my laptop as I ended up with an unsupported WLAN card (Broadcom's, not Linux's fault but the end result is the same). I never managed to get the video pushed out on the projector port. Suspend-to-disk, nope.
This won't improve until the big vendors (Dell, Gateway, HP, IBM) start selling computers where they already have Linux installed and every feature Just Works (tm). Only then will Linux be able to stand shoulder-to-shoulder with M$.
Linksys sells the 802.11g wireless card in my laptop. They used the Broadcom chipset, and the lack of support from Broadcom to enable the driver writers is well-published.
This is one big pain for me in going with Linux on my laptop. Yeah, I got NDIS wrapper to work, but it is painful and I never got the SSID and WEP stuff working right.
I won't recommend Linksys until their Linux support improves. Period.
I saw it as simply a time to separate the wheat from the chaff. Seems to me there were a lot of people who were in IT in 1999-2000 who had no business being there.
While I agree that the industry had too many under-qualified people (and still does), the bubble also hurt extremely talented people who simply were at the wrong company at the wrong time. Friends of mine suffered from months/years of unemployment because of industry conditions and the fact that they were older.
The real tragedy of the bubble is our industry has not learned from it. Another period of high demand will have exactly the same effect.
The last few times I have rented DVDs from the video store, I have had trouble with the physical media. Scratches or whatever. The movie would play until about half way and then just 'hang'. Sometimes I could skip to a scene past the problem and see the rest of the film, but what a pain.
I don't know how to download rips. I suppose I could figure it out but I am just not interested. Yes, I would be willing to pay to have a movie download over night to my Linux server or set-top-box. I can also see value-added services for browsing, previewing and directory searches being worth while.
For me, the biggest issue is what kind of cripple-ware might be included with the downloaded movies. The last DVD I rented (A Perfect Storm) had trouble in the player so I popped it in my laptop. Warner Bros decided to include some special player with the DVD - downloaded and installed automatically and did not work well. I would have liked to just use my existing player, or boot into Linux and use Xine (or whatever).
I have gone to some lengths to try to introduce non-Microsoft products to various clients. I have tried the wedge approach of introducing simple effective tools that work within Windows (e.g. Firefox, Thunderbird, OOo, cygwin, Apache) but have had little success.
It is a serious uphill battle. How can MS collapse with so many organizations 'programmed' to buy their products?
Learn from it, and capture that knowledge somewhere for future procurements. Management 101 - plan what you will do, do it, check the results, take action to correct deficiencies.
Your organization needs a Procurement Management Plan that outlines what types of contracts will be used (FFP, T&M, etc), how estimates will be done, and so on.
As well, invest in training for your management team. You need capabilities in Contract Administration, knowledge of Contract Law and industry best practices.
There are other reasonable ideas in other posts. I am both a contractor and I hire subs on occasion. It can be done well if you take care.
This seems like low short-term risk because you reduce the number of technologies that have to work together, but you incur more long-term risk because of technology churn.
2. Combine libraries
A library implemented in Java/.NET can call a library implemented in.NET/Java using bytecode-IL translators such as IKVM.
Another way is to develop bindings, like we used to do to call C++ libraries from Ada, and such.
3. 'On the wire' integration
This is similar to (2) except that you have more processes.
Using something like CORBA, you can implement a service in, say FORTRAN, that calls the FORTRAN libraries. You can then implement your client in whatever (Java, PERL, C/C++,.NET,...).
There are CORBA/.NET solutions, both OS and commercial, available (Borland Janeva, IIOP.NET,...)
> Have you ever considered that perhaps you're > charging too much for your software?
Thank you - that comment made me laugh.
There is no such thing as an elastic supply/demand curve in software. Even if the developer charged $20, everybody will still use the crack. Laziness, cheapness, whatever. And for their $20, they will expect real effort for support and upgrades. Its a joke.
As has been argued on slashdot so much, the only way to make money from software (nowadays) is through the services you provide on top of it.
Yeah, generating the software in the first place takes a heckova lot of effort, and you won't get paid for it. In effect, you are investing your time to 'write yourself a job' down the road.
I would think that a University should teach CS using as many different languages/environments as practical.
As for seeing people switch from Java to.NET, I suppose there may be a 'bandwagon effect' with all the press that MS has generated, and with many programmers wanting to have.NET on their resume just in case.
Those people switching to.NET now will be just as likely to switch from.NET to the next big thing (maybe.NET++) in five years.
They must have alot of money if they can afford to keep throwing out all that earned value and redo their products with the latest tech.
You mu$t phra$e your propo$al$ in term$ that management under$tand$...
Seriously, though, you need to obtain quantifiable evidence that proves the organization will save money, and how much. Anybody who knows what you are doing will resist you, so watch out...
makes sense, because the government is already 'in the business' of making sure that each of us is uniquely identified (by a SSN or SIN) and our correct address is known (e.g. on our driver's license, voting list).
Now, I also liked the idea of yahoo's DomainKeys proposal, so if the govt could issue me a cert and I could tie it with my DNS info, that would put the existing CA's right out of business...
If you develop a system within a high-speed environment, it is easy to take that network connection for granted.
Recently, I worked at home with a slow connection to work, and noticed the LED on the network hub blinking -alot-. Guess what, I found a place in the code where I was making an RPC call inside a big loop.
Some of these things you never notice unless you test on a slow network...
If your code modifies/extends the GPL code, then your code must also be GPL.
However, if you are just including the GPL component in your system, as in using it as a library/service and making calls to it, then your code does not have to be GPL.
That said, whomever you send your patch to has the right to ask you to provide the GPL'ed source and license. They do not have the right to see your proprietary code.
I would like easier configuration of X. My laptop video card can do RGB out and S-Video out, but you have to change the resolution to make it work.
Shutting down X, hand-editing a config file and restarting X (or starting X with a second config file) is just too hard for most people. The config file is fugly - it is a mish-mash of cryptic options and modelines. I still haven't made my touchpad or keyboard divot work.
There has to be something added to linux that fits between the driver (that makes everything look like a simple file) and the apps that enables us to set configuration for the devices EASILY. It should support non-X and X environments equally. Windoze at least has the 'Display Properties' popup with an "Advanced..." button that somehow knows how to set the options specific to my video card.
Don't get me wrong - I love linux and use it as much as I can. I also use lots of O/S packages in windoze (e.g. cygwin, OOo, Moz). I just think that this gap is the place where the linux environment needs work.
infrastructure to recycle batteries
hydro's capacity to generate and distribute all the electricity required for X households to recharge their car batteries all night
I am not an expert, but I believe that if data is coming to the building, then it will be available to every unit.
It seems to me that the transformer is the natural 'firewall', having very high impedance to high frequency signals. The data has to 'hop' the transformer that connects the apartment to the power grid somehow. As long as there are no impedances between you and your neighbour units, you should all be on the same network. Of course, your neighbour could run a packet sniffer, so you better make sure you are using secure links!
> Care to elaborate on how J2EE doesn't fit > into a hetogeneous environment?
Admittedly, I don't know anything about J2EE Connector, but...
You might want to interface to a scientific application running on a VAX (I worked on one as recently as 1993) implemented in FORTRAN or ADA. Other financial systems are running in COBOL. Telecomms systems are running in C or maybe C++.
With CORBA, you can find an ORB for the language/platform (there are open-source ones available for many languages) and implement a layer of servants that call into the existing system. Then, over the network you can call those services from your client.
Don't think J2EE could do this without lots of JNI and language bindings.
[J2EE] has replaced CORBA as the way to go for
large, distributed enterprise apps
Has J2EE replaced CORBA in those scenarios where either the client or server is NOT written in Java?
One of the facts of life in the enterprise is that it is heterogeneous in terms of platforms, operating systems and (maybe) network technologies. Neither J2EE nor.NET is satisfactory in this environment.
Slashdot Mirror
Awesome - some of us could go into 'Software Law' - a new discipline and second career! Gotta be better than coding...
For me, the biggest threat to Linux is the hardware/driver support.
I gave up on Linux for my laptop as I ended up with an unsupported WLAN card (Broadcom's, not Linux's fault but the end result is the same). I never managed to get the video pushed out on the projector port. Suspend-to-disk, nope.
This won't improve until the big vendors (Dell, Gateway, HP, IBM) start selling computers where they already have Linux installed and every feature Just Works (tm). Only then will Linux be able to stand shoulder-to-shoulder with M$.
This is one big pain for me in going with Linux on my laptop. Yeah, I got NDIS wrapper to work, but it is painful and I never got the SSID and WEP stuff working right.
I won't recommend Linksys until their Linux support improves. Period.
While I agree that the industry had too many under-qualified people (and still does), the bubble also hurt extremely talented people who simply were at the wrong company at the wrong time. Friends of mine suffered from months/years of unemployment because of industry conditions and the fact that they were older.
The real tragedy of the bubble is our industry has not learned from it. Another period of high demand will have exactly the same effect.
The last few times I have rented DVDs from the video store, I have had trouble with the physical media. Scratches or whatever. The movie would play until about half way and then just 'hang'. Sometimes I could skip to a scene past the problem and see the rest of the film, but what a pain.
I don't know how to download rips. I suppose I could figure it out but I am just not interested. Yes, I would be willing to pay to have a movie download over night to my Linux server or set-top-box. I can also see value-added services for browsing, previewing and directory searches being worth while.
For me, the biggest issue is what kind of cripple-ware might be included with the downloaded movies. The last DVD I rented (A Perfect Storm) had trouble in the player so I popped it in my laptop. Warner Bros decided to include some special player with the DVD - downloaded and installed automatically and did not work well. I would have liked to just use my existing player, or boot into Linux and use Xine (or whatever).
rambly post, I know.
I have gone to some lengths to try to introduce non-Microsoft products to various clients. I have tried the wedge approach of introducing simple effective tools that work within Windows (e.g. Firefox, Thunderbird, OOo, cygwin, Apache) but have had little success.
It is a serious uphill battle. How can MS collapse with so many organizations 'programmed' to buy their products?
Learn from it, and capture that knowledge somewhere for future procurements. Management 101 - plan what you will do, do it, check the results, take action to correct deficiencies. Your organization needs a Procurement Management Plan that outlines what types of contracts will be used (FFP, T&M, etc), how estimates will be done, and so on. As well, invest in training for your management team. You need capabilities in Contract Administration, knowledge of Contract Law and industry best practices. There are other reasonable ideas in other posts. I am both a contractor and I hire subs on occasion. It can be done well if you take care.
1. try to do everything in one environment
.NET/Java using bytecode-IL translators such as IKVM.
.NET, ...).
...)
This seems like low short-term risk because you reduce the number of technologies that have to work together, but you incur more long-term risk because of technology churn.
2. Combine libraries
A library implemented in Java/.NET can call a library implemented in
Another way is to develop bindings, like we used to do to call C++ libraries from Ada, and such.
3. 'On the wire' integration
This is similar to (2) except that you have more processes.
Using something like CORBA, you can implement a service in, say FORTRAN, that calls the FORTRAN libraries. You can then implement your client in whatever (Java, PERL, C/C++,
There are CORBA/.NET solutions, both OS and commercial, available (Borland Janeva, IIOP.NET,
> Have you ever considered that perhaps you're
> charging too much for your software?
Thank you - that comment made me laugh.
There is no such thing as an elastic supply/demand curve in software. Even if the developer charged $20, everybody will still use the crack. Laziness, cheapness, whatever. And for their $20, they will expect real effort for support and upgrades. Its a joke.
As has been argued on slashdot so much, the only way to make money from software (nowadays) is through the services you provide on top of it.
Yeah, generating the software in the first place takes a heckova lot of effort, and you won't get paid for it. In effect, you are investing your time to 'write yourself a job' down the road.
I would think that a University should teach CS using as many different languages/environments as practical.
.NET, I suppose there may be a 'bandwagon effect' with all the press that MS has generated, and with many programmers wanting to have .NET on their resume just in case.
.NET now will be just as likely to switch from .NET to the next big thing (maybe .NET++) in five years.
As for seeing people switch from Java to
Those people switching to
They must have alot of money if they can afford to keep throwing out all that earned value and redo their products with the latest tech.
I develop on Win2K, and I use a CLI on Win2K every day.
I just cannot stand the brain-dead command.com environment, so thank god for cygwin!
Correct me if I am wrong, but I think that you can buy/use a TiVo without subscribing to the online directory service.
So, even if the $6 -is- a big deal, you can opt out and still use the TiVo.
You mu$t phra$e your propo$al$ in term$ that management under$tand$ ...
Seriously, though, you need to obtain quantifiable evidence that proves the organization will save money, and how much. Anybody who knows what you are doing will resist you, so watch out...
Any decrease in IE use as seen by your logs is not a true picture.
Some of us Moz/FF/Op users set up our browsers to masquerade as IE, because -some- sites still seem to insist on it...
makes sense, because the government is already 'in the business' of making sure that each of us is uniquely identified (by a SSN or SIN) and our correct address is known (e.g. on our driver's license, voting list).
Now, I also liked the idea of yahoo's DomainKeys proposal, so if the govt could issue me a cert and I could tie it with my DNS info, that would put the existing CA's right out of business...
> which is great for flights to India
:-)
Ha ha - you think they will let you on the plane carrying a bottle of methanol !?!
However, what happens if your place has a fire, gets vandalized, or a burglar takes off with your server(s)?
Good point about the RPC calls.
If you develop a system within a high-speed environment, it is easy to take that network connection for granted.
Recently, I worked at home with a slow connection to work, and noticed the LED on the network hub blinking -alot-. Guess what, I found a place in the code where I was making an RPC call inside a big loop.
Some of these things you never notice unless you test on a slow network...
If your code modifies/extends the GPL code, then your code must also be GPL.
However, if you are just including the GPL component in your system, as in using it as a library/service and making calls to it, then your code does not have to be GPL.
That said, whomever you send your patch to has the right to ask you to provide the GPL'ed source and license. They do not have the right to see your proprietary code.
I would like easier configuration of X. My laptop video card can do RGB out and S-Video out, but you have to change the resolution to make it work.
Shutting down X, hand-editing a config file and restarting X (or starting X with a second config file) is just too hard for most people. The config file is fugly - it is a mish-mash of cryptic options and modelines. I still haven't made my touchpad or keyboard divot work.
There has to be something added to linux that fits between the driver (that makes everything look like a simple file) and the apps that enables us to set configuration for the devices EASILY. It should support non-X and X environments equally. Windoze at least has the 'Display Properties' popup with an "Advanced..." button that somehow knows how to set the options specific to my video card.
Don't get me wrong - I love linux and use it as much as I can. I also use lots of O/S packages in windoze (e.g. cygwin, OOo, Moz). I just think that this gap is the place where the linux environment needs work.
I am not an expert, but I believe that if data is coming to the building, then it will be available to every unit.
It seems to me that the transformer is the natural 'firewall', having very high impedance to high frequency signals. The data has to 'hop' the transformer that connects the apartment to the power grid somehow. As long as there are no impedances between you and your neighbour units, you should all be on the same network. Of course, your neighbour could run a packet sniffer, so you better make sure you are using secure links!
> It doesn't sound like you have much
> of a clue about J2EE
Not very tactfully put AC, but yes you are right - my J2EE experience is limited.
> We interface our payroll java production
> code with COBOL. Not much of a problem.
How? Using JNI->C->COBOL or another mechanism?
> Care to elaborate on how J2EE doesn't fit
> into a hetogeneous environment?
Admittedly, I don't know anything about J2EE Connector, but...
You might want to interface to a scientific application running on a VAX (I worked on one as recently as 1993) implemented in FORTRAN or ADA. Other financial systems are running in COBOL. Telecomms systems are running in C or maybe C++.
With CORBA, you can find an ORB for the language/platform (there are open-source ones available for many languages) and implement a layer of servants that call into the existing system. Then, over the network you can call those services from your client.
Don't think J2EE could do this without lots of JNI and language bindings.
[J2EE] has replaced CORBA as the way to go for
.NET is satisfactory in this environment.
large, distributed enterprise apps
Has J2EE replaced CORBA in those scenarios where either the client or server is NOT written in Java?
One of the facts of life in the enterprise is that it is heterogeneous in terms of platforms, operating systems and (maybe) network technologies. Neither J2EE nor