Remember Mono is only really suitable for writing Suse apps. If you didn't get it from Novell, you're liable to be sued as the patent protection only applies to the deal they have with Novell, not anyone else.
It still doesn't have WPF or LINQ support does it now?
To anyone thinking about going C#, I woudl strongly consider using VB.NET instead. Yup, I know its VB, but think about this: it and C# both turn into the same code, they both have roughly the same syntax - to the point where you can say C# is VB with curly brackets!
VB is a lot easier to use, they have features in the tools that are much nicer than C#s, and they have a few features that C# doesn't have like conditional exceptions (eg. catch x as exception when x.code = 10)
It also have the advantage that you don't get so mind-muddled swapping between C/C++ and C# as it tries to get credibility by looking like C++ code. Until MS produces a first-class Python language for.NET, really go with VB. You can also leverage your skills to writing wscript code them too (in VBA). And, its so blinking obvious you don't need to learn anything, the IDE helps you along.
The only problem is that people will think you're not a real programmer, but if you're writing.net apps in either, you're not anyway:)
I'd say there might have been a performance difference in the languages, back when a function pointer lookup was considered something to optimise away.. but I think the biggest problem was the people. The people who suddenly got hold of C++ and decided they must use objects, and did so...
Reminds me of the Windows Explorer team when they were writing it for Windows 95. Suddenly they were told the platform had threads, so they should use them in explorer. End result was a thousand threads running all the time, one for each little expand button in the tree view, etc etc. They re-wrote large chunks to make it perform badly (ie the previous version didn't, as you'd expect).
The problem there wasn't that threads are bad, but that overuse and poor design cosiderations are the problem. The same applies to misused features of C++ (and any language), like an object for everything.
CORRECT program will release memory when it is no longer used. If it desn't do that, especially in C, then something is terribly wrong with it, and it should not be used until it's fixed.
there, fixed that for you. You can even replace "C" with any language you care to mention, including the ones with garbage collectors and the statement is still true.
true, in fact the earliest C++ "compiler" was a translator tool that turned C++ into C code to be compiled by a C compiler (ie before there were C++ compilers). Look up cfront for details.
Personally, I never code C anymore, I always use a C++ compiler with handy stuff like the STL.
As long as you export functions wrapped with "extern C" no-one using your code will ever need know you've done that. (ok, this assumes you're working with shared libraries/dlls and not re-using code directly).
Sometimes its worse than that - you have the old product being actively maintained for existing customers, but no new ones want to buy it and they cannot buy the new one until it's finished. You also have your best (hmm, or those who think they're the best) writing (or more often over-engineering) the new code all being paid salary that is not recouped by sales.
Net result for one company I worked for that did this: a whisker within bankruptcy and loads of redundancies.
When I was a young developer, I would be thrilled with the thought of rewriting code in shiny new tools, and remove the crufty old, and make it more usable and faster and etc. Now I'm an old developer I know better that there is much much more to developing products than coding them up. (and I also know new != better).
A large codebase takes so much time and energy for a newbie to understand that often it is more effective, and nearly as quick... and a lot more buggy.
A re-write from scratch is nearly always the wrong approach. Not only is it the same coders who created the old POS going to be responsible for the "shiny new cleanly designed codebase" (ie thew new POS) but you lose all the experience of working with the old one, often at a time when you still have to maintain it while doing the rewrite.
in standardizing our low-end devices on a single mobile OS, though we have not finalized our plans to do so." '
So, you'll get probably get a crippled/slow device
no, it means that the phones without touchscreen, music libraries, 20mp cameras, little ram.. you know the phones people want to make calls and text, they'll be operated by a common platform. There won't be anything crippled about them, they'll be designed for a certain market segment.
Modern CPU's have 8+ Mega Bytes of L2/L3 cache on chip so RAM is only a problem when your working set it larger than that.
Unfortunately, most modern apps require far more working set than that! The crappy.NET app I use at work had a working set of 700MB today.
The other issue is that HPC applications generally require small amounts of processing on lots and lots of snippets of data - ie highly parallel processing. This means that memory bandwidth is a very significant bottleneck.
you can have lot's of cores and a tiny cache like GPU's
Incidentally GPUs have a lot of cache - my Graphics card has 512Mb RAM.
I'd say that bugtracking, ticket tracking OSS software is one area where OSS far outperforms the commercial stuff. We use Siebel for our corporate stuff, and... I hate it with a true passion. We also use Mantis for internal bug tracking, so far 3 groups have decided that they want to continue using it, or move to using it.
Same applies to Remedy and Clarify and.. all the others.
That does not necessarily apply to other areas, I find that there is some OSS software that is brilliant, and a ton that is poop, but if you can find the good stuff you'll find it is generally better than the commercial equivalent.
I'm not sure the majority of people want to "beat Microsoft" (well, ok, this is/.) but Microsoft has been set as a target - or milestone, or goal to reach. So getting the the market share of MS would not be about beating MS, but of making an OS that achieves enough usability and functionality that people want and start using it.
In the industry, success is measured in market share, if Linux could get quarter the share as Windows has, then I'd say that it had reached the mainstream. Wanting and striving for that goal doesn't have to disparage Windows at all.
Some companies have no choice. For example, if you have a Microsoft gold partnership, you get lots of 'free' software (sortof, it costs for the packs in the first place, but its lots cheaper than buying the sw individually). However, if you have such an agreement you'll find you're licenced only for the latest version.
We're looking at a choice here - upgrade to Vista/Office 2007, or buy the old software. In a way we have Microsoft on one side, a rock on the other and my IT manager's danglies in the middle.
I'll meet you half way, I agree C++ is far faster than anything I've been paid to code in. Now you come the other half of the way and maybe we'll have a discussion here where useful information is exchanged.
and this is the point where great things get done. I've always coded my best scalable, maintainable, and generally expandable and flexible apps when I de-couple the GUI from the Application (and often the Data layers too)(without being an architecture nazi about stored procedures and client-side validation).
So I really think the best way to do applications is to write them like this, whether its a web front end to a VB GUI front end to the same back-end code. The back end is written in C++ for the speed and the power, the front end in whatever looks nice and/or is a good language for manipulating the data that is managed by the back-end code. Once we did write GUIs in VB, now customers want the same applications but with a web GUI. God knows what they'll want next - mobile front ends probably.
You don't have to think that C++ is the ultimate tool, but it turns out repeatedly that it is the right tool for those back-end applications.
I can say this is the model that directly affects us. I work for a large US corporate who sells enterprise software to governments and similar. We've been beaten in bids by our #1 competitor for a while, it turns out they are offering the software for free, as long as the buyer takes on a services contract (think outsourced IT type thing) that would support the software and hardware required to run it.
We sell the software and let the buyer decide how to get support for hardware and other IT systems (we provide serious support for our software only, not stuff like Windows and email etc). So far, we've lost every bid. Unfortunately, our US overlords won't let us change our terms.
So looking at this from a FOSS POV, it is a model that can work - give your software away for free, and then go and sell your consultancy and support services to corporates who buy it off you. You should be providing them with support services to get it installed and configured, not just "if it breaks we'll look at it for you" and "bugfix" support.
Yeah, that means you have to work all the time, you can't just make something then lie back and see the money rolling in, but that old 'make loads of easy money without having to work" is a paradigm that died last year.
We need something similar for society. Every year, your employer, doctor, family and friends should send an appraisal to the government. The bottom 10% of society could be put to sleep.
Yeah, but after all the politicians have been put to sleep, who'll decide who comprises the remaining 8%?
You mean the AC should have an auction to decide which of the co-workers gets to pull the trigger on his behalf? From what the AC posted, I imagine the final bid would be quite large!
(PS. I don't think the AC was simply fantasising about his miserable life and trying to troll us)
I thought the generally accepted practice for MS is to first blame the video driver, and then blame the printer driver. *then* they might look at the problem:)
Mind you, I agree with MS here, the biggest problem with the browser is the add-ins.. ones like SmileyCentral, AdsULike, PhishingToolbar, AntiVirusCheckPro, and NoSpamHonestNoReally.
if you want old computers, go down the recycling centre. you'd be surprised how many good PCs get thrown out.
I'm not sure if they'll let you take them (though the bloke at my local tip was happy for me to have a few bits) but if not, you can hang around and ask people who are bringing their old PCs to throw away to give them to your cause (or get a big poster up)
its a bit of a grey area how much dynamic the CLR has. C# currently has support for "var" variables (the 'work out the type at runtime' variable type), and extension methods (which are definitely a dynamic language feature).
I'm sure the DLR is coming, but it might just be branded as a new version of the CLR (perhaps one that is not released to the standards body this time)
PS. reflection achieves good performance?!?!?!? first time I've heard that!:)
yep, the Novell/Microsoft agreement only covers stuff downloaded directly from Novell. So, if you do apt-get moonlight on Debian, you may get sued or end up having to pay royalties.:)
Slashdot Mirror
hear hear.
Remember Mono is only really suitable for writing Suse apps. If you didn't get it from Novell, you're liable to be sued as the patent protection only applies to the deal they have with Novell, not anyone else.
It still doesn't have WPF or LINQ support does it now?
To anyone thinking about going C#, I woudl strongly consider using VB.NET instead. Yup, I know its VB, but think about this: it and C# both turn into the same code, they both have roughly the same syntax - to the point where you can say C# is VB with curly brackets!
VB is a lot easier to use, they have features in the tools that are much nicer than C#s, and they have a few features that C# doesn't have like conditional exceptions (eg. catch x as exception when x.code = 10)
It also have the advantage that you don't get so mind-muddled swapping between C/C++ and C# as it tries to get credibility by looking like C++ code. Until MS produces a first-class Python language for .NET, really go with VB. You can also leverage your skills to writing wscript code them too (in VBA). And, its so blinking obvious you don't need to learn anything, the IDE helps you along.
The only problem is that people will think you're not a real programmer, but if you're writing .net apps in either, you're not anyway :)
I'd say there might have been a performance difference in the languages, back when a function pointer lookup was considered something to optimise away.. but I think the biggest problem was the people. The people who suddenly got hold of C++ and decided they must use objects, and did so...
Reminds me of the Windows Explorer team when they were writing it for Windows 95. Suddenly they were told the platform had threads, so they should use them in explorer. End result was a thousand threads running all the time, one for each little expand button in the tree view, etc etc. They re-wrote large chunks to make it perform badly (ie the previous version didn't, as you'd expect).
The problem there wasn't that threads are bad, but that overuse and poor design cosiderations are the problem. The same applies to misused features of C++ (and any language), like an object for everything.
CORRECT program will release memory when it is no longer used. If it desn't do that, especially in C, then something is terribly wrong with it, and it should not be used until it's fixed.
there, fixed that for you. You can even replace "C" with any language you care to mention, including the ones with garbage collectors and the statement is still true.
true, in fact the earliest C++ "compiler" was a translator tool that turned C++ into C code to be compiled by a C compiler (ie before there were C++ compilers). Look up cfront for details.
Personally, I never code C anymore, I always use a C++ compiler with handy stuff like the STL.
As long as you export functions wrapped with "extern C" no-one using your code will ever need know you've done that. (ok, this assumes you're working with shared libraries/dlls and not re-using code directly).
Sometimes its worse than that - you have the old product being actively maintained for existing customers, but no new ones want to buy it and they cannot buy the new one until it's finished. You also have your best (hmm, or those who think they're the best) writing (or more often over-engineering) the new code all being paid salary that is not recouped by sales.
Net result for one company I worked for that did this: a whisker within bankruptcy and loads of redundancies.
When I was a young developer, I would be thrilled with the thought of rewriting code in shiny new tools, and remove the crufty old, and make it more usable and faster and etc. Now I'm an old developer I know better that there is much much more to developing products than coding them up. (and I also know new != better).
A large codebase takes so much time and energy for a newbie to understand that often it is more effective, and nearly as quick ... and a lot more buggy.
A re-write from scratch is nearly always the wrong approach. Not only is it the same coders who created the old POS going to be responsible for the "shiny new cleanly designed codebase" (ie thew new POS) but you lose all the experience of working with the old one, often at a time when you still have to maintain it while doing the rewrite.
in standardizing our low-end devices on a single mobile OS, though we have not finalized our plans to do so." '
So, you'll get probably get a crippled/slow device
no, it means that the phones without touchscreen, music libraries, 20mp cameras, little ram.. you know the phones people want to make calls and text, they'll be operated by a common platform. There won't be anything crippled about them, they'll be designed for a certain market segment.
Modern CPU's have 8+ Mega Bytes of L2/L3 cache on chip so RAM is only a problem when your working set it larger than that.
Unfortunately, most modern apps require far more working set than that! The crappy .NET app I use at work had a working set of 700MB today.
The other issue is that HPC applications generally require small amounts of processing on lots and lots of snippets of data - ie highly parallel processing. This means that memory bandwidth is a very significant bottleneck.
you can have lot's of cores and a tiny cache like GPU's
Incidentally GPUs have a lot of cache - my Graphics card has 512Mb RAM.
I'd wager if you put 8x the minimum required memory in your Vista box, you'll get all the performance you could wish for.
Unfortunately I don't have that many RAM slots in my motherboard, so its kind of a rhetorical solution :(
I'd say that bugtracking, ticket tracking OSS software is one area where OSS far outperforms the commercial stuff. We use Siebel for our corporate stuff, and ... I hate it with a true passion. We also use Mantis for internal bug tracking, so far 3 groups have decided that they want to continue using it, or move to using it.
Same applies to Remedy and Clarify and .. all the others.
That does not necessarily apply to other areas, I find that there is some OSS software that is brilliant, and a ton that is poop, but if you can find the good stuff you'll find it is generally better than the commercial equivalent.
I'm not sure the majority of people want to "beat Microsoft" (well, ok, this is /.) but Microsoft has been set as a target - or milestone, or goal to reach. So getting the the market share of MS would not be about beating MS, but of making an OS that achieves enough usability and functionality that people want and start using it.
In the industry, success is measured in market share, if Linux could get quarter the share as Windows has, then I'd say that it had reached the mainstream. Wanting and striving for that goal doesn't have to disparage Windows at all.
if that means that the driver disk you got with your all-in-one scanner-copier-coffee machine doesn't work after 2 years, so be it.
yeah, can we get back on-topic, we're discussing Linux here, not Windows. :)
Some companies have no choice. For example, if you have a Microsoft gold partnership, you get lots of 'free' software (sortof, it costs for the packs in the first place, but its lots cheaper than buying the sw individually). However, if you have such an agreement you'll find you're licenced only for the latest version.
We're looking at a choice here - upgrade to Vista/Office 2007, or buy the old software. In a way we have Microsoft on one side, a rock on the other and my IT manager's danglies in the middle.
I'll meet you half way, I agree C++ is far faster than anything I've been paid to code in. Now you come the other half of the way and maybe we'll have a discussion here where useful information is exchanged.
and this is the point where great things get done. I've always coded my best scalable, maintainable, and generally expandable and flexible apps when I de-couple the GUI from the Application (and often the Data layers too)(without being an architecture nazi about stored procedures and client-side validation).
So I really think the best way to do applications is to write them like this, whether its a web front end to a VB GUI front end to the same back-end code. The back end is written in C++ for the speed and the power, the front end in whatever looks nice and/or is a good language for manipulating the data that is managed by the back-end code. Once we did write GUIs in VB, now customers want the same applications but with a web GUI. God knows what they'll want next - mobile front ends probably.
You don't have to think that C++ is the ultimate tool, but it turns out repeatedly that it is the right tool for those back-end applications.
but does it aggregate data from "5-10 servers" and display them in a way the boss can easily access, like on a web page?
I can say this is the model that directly affects us. I work for a large US corporate who sells enterprise software to governments and similar. We've been beaten in bids by our #1 competitor for a while, it turns out they are offering the software for free, as long as the buyer takes on a services contract (think outsourced IT type thing) that would support the software and hardware required to run it.
We sell the software and let the buyer decide how to get support for hardware and other IT systems (we provide serious support for our software only, not stuff like Windows and email etc). So far, we've lost every bid. Unfortunately, our US overlords won't let us change our terms.
So looking at this from a FOSS POV, it is a model that can work - give your software away for free, and then go and sell your consultancy and support services to corporates who buy it off you. You should be providing them with support services to get it installed and configured, not just "if it breaks we'll look at it for you" and "bugfix" support.
Yeah, that means you have to work all the time, you can't just make something then lie back and see the money rolling in, but that old 'make loads of easy money without having to work" is a paradigm that died last year.
We need something similar for society. Every year, your employer, doctor, family and friends should send an appraisal to the government. The bottom 10% of society could be put to sleep.
Yeah, but after all the politicians have been put to sleep, who'll decide who comprises the remaining 8%?
You mean the AC should have an auction to decide which of the co-workers gets to pull the trigger on his behalf? From what the AC posted, I imagine the final bid would be quite large!
(PS. I don't think the AC was simply fantasising about his miserable life and trying to troll us)
I thought the generally accepted practice for MS is to first blame the video driver, and then blame the printer driver. *then* they might look at the problem :)
Mind you, I agree with MS here, the biggest problem with the browser is the add-ins.. ones like SmileyCentral, AdsULike, PhishingToolbar, AntiVirusCheckPro, and NoSpamHonestNoReally.
bmwm3nut is right, "tip" is a colloquial term for "place you take stuff the local refuse collectors wouldn't take".
if you want old computers, go down the recycling centre. you'd be surprised how many good PCs get thrown out.
I'm not sure if they'll let you take them (though the bloke at my local tip was happy for me to have a few bits) but if not, you can hang around and ask people who are bringing their old PCs to throw away to give them to your cause (or get a big poster up)
its a bit of a grey area how much dynamic the CLR has. C# currently has support for "var" variables (the 'work out the type at runtime' variable type), and extension methods (which are definitely a dynamic language feature).
I'm sure the DLR is coming, but it might just be branded as a new version of the CLR (perhaps one that is not released to the standards body this time)
PS. reflection achieves good performance?!?!?!? first time I've heard that! :)
yep, the Novell/Microsoft agreement only covers stuff downloaded directly from Novell. So, if you do apt-get moonlight on Debian, you may get sued or end up having to pay royalties. :)