This is more akin to using a debugger than declarative programming.Many people ahve this misconception of declarative programming as just writing a specification then the prgram working out what to do whereas in reality it means having to redfine every algorithm in terms of recursion. I can assure you that writing in prolog certainly does nothing to improve comprehension!
I find Outlook a terribly annoying email client to use, and would prefer to use my own. Im forced to use Lotus Notes and pray one day to be able to return to using Outlook!
I dont see the problem with J2EE, it saves work not makes it. You write your app to conform to J2EE architecture which requires some reading, research and thinking. But then in return you can swap out pretty much any component to other vendors or in some cases other technologies.
Want to change all your comms to go from email to nntp, no problem since you coded using the J2EE interfaces you just need to change 1 config file to point to the right jar.
Want to log to a remote syslog server instead of flat files (or the windows event log), no problem, change one config file in your log4j setup.
Now admittedly the time investment learning J2EE architecture and the associated frameworks / vendor specific components takes a while but the resulting product is remarkably adaptable and fault tolerant. True one dynamic web site, serving data from one database may not jsutify it but this stuff is about linking together many disparate systems.
Not quite since the error rate quoted is for any error, however being marked dead within your lifetime requires the mistake to correlate with a living person.
For example you haven't taken into account when someone dies twice as a result of a mistake, or they kill of people like Jxo Blfggs (who presumably don't exist)
But the thing is it will almost certainly save time, and not jsut in the long run. Straight away by giving an auditable trail of responsibility to the reviewer as well as the author you end up with less commits which break the build (since the reviewer should be building the commit in a sandbox) which can stop other developers becoming blocked and ensure there is always a stable codebase should you need to release
Well this is clearly a process problem. If you can't commit without a peer review, which includes documentation standards, then none of the uncommented and undocumented destruction ever occurs.
Rather than going back and using "academic types" to polish your code why not put the effort into the design, if this is right the code practically writes itself.
Except that if the programmer is coming from a proprietary shop they may not even be *allowed* to contribute to open source projects. (I know Im not without management authorization and the company I work at *definitely* hires some superstars). Sure if they're out there in Open Source its a reasonable thing to do but I wouldnt use it as point of filtering.
More to the point, networks and cable stations do this regularly to air movies on TV. Rather than prohibit *other* people from editing movies, movie studios should do what they do for networks - provide a license to edit the movie. Yes exactly. However creating a non-authorized derivitive work is theft, umm I mean piracy, I mean *sharing*.
I think management probably do grasp that lines of code arent necesarily proportional to the amount of work, however they are probably wanting to make sure the the developer are doing *some* work.
Personally I would gather 1/(no. slash dot comments):D
So how do you debug your application by grepping through the callstack and opening a buffer each time? Bleurgh. Vi for quick edits over ssh (since its the only thing that is instaled EVERYWHERE), eclipse for java/perl/scripting/xml etc, VisualStudio for c++/c, SlickEdit for reading trace files (its search / large file handling is great)
Round up 7 or 8 CDs worth of software. Many of them needing codes and activation (or cracks). Install these one by one.
I'd create a images and force matching computers. But, I (like many small business types) am dealing with only 15 or 20 computers, every single one of which is unique. Hope your not cracking that software on business machines..thats...unwise
I agree with the gp, these are all simple to configure and certainly no more arcane than some of linux's man pages. Incidently about 80% of my linux use is via ssh from my windows box because I find certain apps on windows a lot less fuss and nonsense to use than on linux.
At the end of the day linux doesnt spread because most computer users dont care what operating system they are running. They use the computer as a tool to do something else and as long as they have something that is good *enough* what motivitaion is there for them to switch?
Surely in linux airlines you are fired through a series of pipes
Re:XML is a fad, STEP is the future
on
The Future of XML
·
· Score: 1
Which isnt really very human editable is it? It wont take long before you get:
#23413=SOMETHING(#23, 'somestuff' #2345, 'somemorestuff', #9987976)
At least in XML you have a kind of proximity effect, so yes if your searching for a particular element you could at worst case have to scan the hwole file. But when you instantiate an element you dont have to jump around as much.
Also your lines arent fixed width so seeking to #9987976 means scanning through for 9987975 newline characters.
Re:"How will you use XML in years to come?"
on
The Future of XML
·
· Score: 1
he benefit of having a national ID card on the other hand is, that there's only a small number of documents used commonly and if you have one, you are identified. No more 'Bring 3 types of ID' stuff. You have your driving license, your passport or your ID card, you are set. If those are good enough for the police, they are good enough for everyone else too (eg banks, insurances, airlines). Most things require proof of ID and 2 forms of proof of address.
Actually, Linux as RMS uses it really is pretty much just GNU/Linux. I understand he doesn't use X or anything that requires a GUI, just EMACS, GNU Screen and BASH I dont understand why this kind of thing is idolised. This guy is probably THE worst ambassador for free software / open source adoption, certainly in the corporate world.
Is he even still an active contributer does he just tour the world giving embarressing rants?
Slashdot Mirror
This is more akin to using a debugger than declarative programming.Many people ahve this misconception of declarative programming as just writing a specification then the prgram working out what to do whereas in reality it means having to redfine every algorithm in terms of recursion. I can assure you that writing in prolog certainly does nothing to improve comprehension!
The following is a screenshot from prolog-Quake:
?- No.
I dont see the problem with J2EE, it saves work not makes it. You write your app to conform to J2EE architecture which requires some reading, research and thinking. But then in return you can swap out pretty much any component to other vendors or in some cases other technologies.
Want to change all your comms to go from email to nntp, no problem since you coded using the J2EE interfaces you just need to change 1 config file to point to the right jar.
Want to log to a remote syslog server instead of flat files (or the windows event log), no problem, change one config file in your log4j setup.
Now admittedly the time investment learning J2EE architecture and the associated frameworks / vendor specific components takes a while but the resulting product is remarkably adaptable and fault tolerant. True one dynamic web site, serving data from one database may not jsutify it but this stuff is about linking together many disparate systems.
So say you ground-dweller, plus the wi-fi extends for 1.6km up here.
Why stop there, I propose we make ALL numbers period self checksumming. The axiom of closure is over rated anyway.
Not quite since the error rate quoted is for any error, however being marked dead within your lifetime requires the mistake to correlate with a living person.
For example you haven't taken into account when someone dies twice as a result of a mistake, or they kill of people like Jxo Blfggs (who presumably don't exist)
Precisely, which is why the best we can do is forward date your death certificate in 6 monthly rolling intervals. Sorry
But the thing is it will almost certainly save time, and not jsut in the long run. Straight away by giving an auditable trail of responsibility to the reviewer as well as the author you end up with less commits which break the build (since the reviewer should be building the commit in a sandbox) which can stop other developers becoming blocked and ensure there is always a stable codebase should you need to release
But only if they're statically linked.
Well this is clearly a process problem. If you can't commit without a peer review, which includes documentation standards, then none of the uncommented and undocumented destruction ever occurs.
Rather than going back and using "academic types" to polish your code why not put the effort into the design, if this is right the code practically writes itself.
Except that if the programmer is coming from a proprietary shop they may not even be *allowed* to contribute to open source projects. (I know Im not without management authorization and the company I work at *definitely* hires some superstars). Sure if they're out there in Open Source its a reasonable thing to do but I wouldnt use it as point of filtering.
In my RSS reader it totally looked like $28. I was totally ready to get in on the bidding myself!
I think management probably do grasp that lines of code arent necesarily proportional to the amount of work, however they are probably wanting to make sure the the developer are doing *some* work. Personally I would gather 1/(no. slash dot comments) :D
If you think Visual Studio is a glorified text editor then you clearly havent used it enough.
So how do you debug your application by grepping through the callstack and opening a buffer each time? Bleurgh. Vi for quick edits over ssh (since its the only thing that is instaled EVERYWHERE), eclipse for java/perl/scripting/xml etc, VisualStudio for c++/c, SlickEdit for reading trace files (its search / large file handling is great)
Except that shareholders are typically promised by CEOs double digit growth, which forces large customers to expand into other markets.
I'd create a images and force matching computers. But, I (like many small business types) am dealing with only 15 or 20 computers, every single one of which is unique. Hope your not cracking that software on business machines..thats...unwise
I agree with the gp, these are all simple to configure and certainly no more arcane than some of linux's man pages. Incidently about 80% of my linux use is via ssh from my windows box because I find certain apps on windows a lot less fuss and nonsense to use than on linux.
At the end of the day linux doesnt spread because most computer users dont care what operating system they are running. They use the computer as a tool to do something else and as long as they have something that is good *enough* what motivitaion is there for them to switch?
Surely in linux airlines you are fired through a series of pipes
Which isnt really very human editable is it? It wont take long before you get:
#23413=SOMETHING(#23, 'somestuff' #2345, 'somemorestuff', #9987976)
At least in XML you have a kind of proximity effect, so yes if your searching for a particular element you could at worst case have to scan the hwole file. But when you instantiate an element you dont have to jump around as much.
Also your lines arent fixed width so seeking to #9987976 means scanning through for 9987975 newline characters.
Only if your XML Parser doesnt support W3C XML Schema Part 2.
My sugar cubes won't fit through the pipe.
Is he even still an active contributer does he just tour the world giving embarressing rants?