By case I meant Computer Aided Software Engineering, e.g. the system within Microsoft Visual Studio where C++ code (using MFC, is generated for you by the "wizard"). So yes, I guess something like C++ builder.
Any CASE tools do tend to be detesbable. However, more so than VM based languages? That messy generated code is, at least, accessible, the messy code in a VM is just your environment, so live in it.
It is the confusion about XML, the hype, and all the metalanguages created using XML that takes up so much extra paper. XML itself could be explained in that space, but how could you leave out XSTL? And don't you want to talk about RSS and SMIL and SOAP, etc. etc.
That is, some have heard many false promises regarding XML (which they should not have believed, imho, anyway, but still). XML does nothing for you, in the same way that the Heirarchical File System does nothing for you. The HFS allows you to use directories to organize things, but as we know, most file systems are still abysmally disorganized. Damn HFS!
XML allows you to have forward, backward, and sideways compatibility in file formats, it allows you, via XSLT, a path for just-in-time compatibility. But it doesn't actually have any way to promote these things... it just makes them possible.
> The app you write in a couple days the VB programmer can toss out after lunch.
I have not found this to be the case, ever. Claimed a lot, but not true, in my experience.
Especially since it is the CASE tools that accomplish this for VB, and they "work" for C++ too. Of course CASE tools introduce bloat and inefficiency, but the CASE generated C++ is still more efficient than the VB equiv.
Ok, done Sun's way might be the same as "done right", but then doing VB "right" is still a dubious achievment.
And look the whole industry is VM hungry (when considering all the other scripting engines), with Sun's blessing. I'm not sure the herd is right on this one.
And although Java is a resounding success on the server... has it made Sun any money?
Thing is, people are calling this OS after the kernel, they can. Windows is named after it's windows manager. Many people call the OS they are using, "RedHat".
FSF could come out with a distro and call it GNU.
Further, the idea that GNU doesn't get credit is'nt really fair, how long can you use Linux without reading GNU this and GNU that?
It's a test of freedom, and the software is winning.
I realize it looked like I was boasting, but I was not, there is nothing particularly special about working in the game industry except to support my comments about where the game industry was in terms of code reuse.
It seemed boastful only because of your personal insecurity and --- what... your name really is Pyrrho? oh, sorry.
as far as getting you started, I've read some of your exposition on this.
I don't say that OO does it better, I think this book, Design Patterns, which happens to be a part of OOD, deals with the subject matter appropriates. It does not advocate an extreme, it advocates and understanding of the implications of different kinds of relationships.
And the reason why is exactly what you mention, understanding the implications of various kinds of relationships (aka coupling between components) is what's important, so that in each case you can decide what you need. Tight coupling is good in some cases, loose in others. And of course, there is gradiant from one extreme to the other.
On the other hand, I do think in terms of OOD, but not in the way you think of OOD. My OOD is much more pragmatic than the OOD that seems to have become a pet peeve of yours.
Believe me, I know what irks you in the sense that there certainly is a class of OOProgrammer that give OOD a bad name. That is, bad designs defended as good because they are object oriented.
there is paradigm bias... they understand data and functions as coupled, there is an OOD bias, but not bigotry.
Just as there is a bit of a C++ bias, but that's just a pragmatic factor.
The way Design Patterns handled coupling issues is very interesting. They don't say something like sickOOPs say, "decouple everything" "abstract everything", etc., but they tell you ways to decouple, and give insight on the benefits and costs of certain kinds of dependencies from a purely logical point of view.
Why not a C+SQL pattern books... well, there is room there for such a book. Problem: the "patterns" area of books is a sick wasteland, Design Patterns isn't just the shining star, it's practically the only decent example (though no doubt someone can point me to some exception, this is imho, ymmv).
by the way, the waterfall method is long since discredited, so if the logician doesn't have anything to do with the design of his code then he is at a bad company, a company that would make even interesting projects boring and pointless. Did you consider that?
Re:Separating Content from Presentation a Good Thi
on
Office 2003 and XML
·
· Score: 1
you are supposed to separate the formatting/style... NOT DISCARD IT!
The internet and shrink wrap software business mayhem have ruined the sensibilities of many programmers to the point they don't realize...
When people pay high $$$ for an "engineer", they WANT to be told how to do it right! And even cheap tech labor is expensive.
But after working for 23 year olds that just got 50G$ from some VC, they think everybody is an impetuous youth that wants the impossible, and if they can't have it, they would prefer a quick piece of crud to a tool that actually works!
When I read your sig, or the version that changes it to "slashdot" (don't know if that's you or not), I want to laugh, I get it... but everytime my compulsive gaurdian geek pops into view on my shoulder and says, "A million is no where near infinity... compared to infinity, a million is no more than 1."
well, tablizer, having read your arguments about OOP your comment is not suprising.
But consider... Design Patterns is a logic book, it's about logical relations, and coupling, loose and tight, about references, linked, embeded, direct and indirect.
They are not templates. That is, you can impliment a pattern using SQL and C if you so desire! The implimentation is left for your (or some vender that wants to serve your needs), and that's the way it should be.
Actually Design Patterns (not talking about other pattern books, just Design Patterns) is probably the best book for you in particular as it doesn't say that OOD is the be all end all. In fact, when they give their example of a OO parser, they specifically say this is not the best way to build a parser, there are more efficient and standard structured approaches, and the pattern is presented to investigate the logic and functional relationships involved.
In fact, games are no longer this way, or at least, it's not the profitable way.
Games need to use libraries, need to have engines that can be brought forward instead of thrown out.
In the old days the need to optimize for slow machines was a big pressure, you didn't have a general purpose 3D engine, you had something that produced a picture and underneath was really a bunch of special optimization tricks you didn't want to bring forward.
Now, the natural progression of reuse is entering the game industry. Look how well the Sims is doing, partially because they can release the Date Pack (or whatever) and improve their system instead of replace it.
and yes, I worked in the game industry (network focussed) for over 6 years.
The internet bubble has burst. Time to market over quality was a part of that bubble.
message over.
PS: to be serious, you do have a point, and time to market will always be important, but it's not the "time it takes me to have something I can put in the market" it will return to being "the time it takes to put together something that WORKS, and get it to market."
(1) I think there is more tedium than excitment protecting a village from bandits.
(2) All software projects can be interesting: if you are a logician. If you love logic. All these system are interesting to design and construct (less interesting, imho, to maintain and operate!)
Of course, I'm a bit full of it as I've always gravitate to parts of the industry that interest me, because it does help one motivate oneself, not so much for projects that are more logically interesting, but to stay motivated through political storms and other meatspace crud.
They've had to "uncouple" IE. Now they're being asked to uncouple Media Player. What's next? Uncoupling the calculator? The start button? Command prompt? Following this line of thinking ad absurdum, what exactly is Microsoft allowed to package with Windows? Sheesh!
Replaceable componentry is usually advisable based on engineering principle, but in this case legal systems are imposing a component based approach, attempting to protect their free markets.
Slashdot Mirror
By case I meant Computer Aided Software Engineering, e.g. the system within Microsoft Visual Studio where C++ code (using MFC, is generated for you by the "wizard"). So yes, I guess something like C++ builder.
Any CASE tools do tend to be detesbable. However, more so than VM based languages? That messy generated code is, at least, accessible, the messy code in a VM is just your environment, so live in it.
>But XML is so cumbersome otherwise that almost all of it will be either machine generated or edited in special editors.
good thing machines never corrupt files or we would want to validate machine output as well.
It is the confusion about XML, the hype, and all the metalanguages created using XML that takes up so much extra paper. XML itself could be explained in that space, but how could you leave out XSTL? And don't you want to talk about RSS and SMIL and SOAP, etc. etc.
That is, some have heard many false promises regarding XML (which they should not have believed, imho, anyway, but still). XML does nothing for you, in the same way that the Heirarchical File System does nothing for you. The HFS allows you to use directories to organize things, but as we know, most file systems are still abysmally disorganized. Damn HFS!
XML allows you to have forward, backward, and sideways compatibility in file formats, it allows you, via XSLT, a path for just-in-time compatibility. But it doesn't actually have any way to promote these things... it just makes them possible.
but what the heck does he mean that "callback are non-idiomatic". I believe that callbacks are an idiom.
And one many of us are quite used to! Using SAX you don't have to load the entire document.
> The app you write in a couple days the VB programmer can toss out after lunch.
I have not found this to be the case, ever. Claimed a lot, but not true, in my experience.
Especially since it is the CASE tools that accomplish this for VB, and they "work" for C++ too. Of course CASE tools introduce bloat and inefficiency, but the CASE generated C++ is still more efficient than the VB equiv.
imnsho
... does this theory of mine make sense?
Ok, done Sun's way might be the same as "done right", but then doing VB "right" is still a dubious achievment.
And look the whole industry is VM hungry (when considering all the other scripting engines), with Sun's blessing. I'm not sure the herd is right on this one.
And although Java is a resounding success on the server... has it made Sun any money?
bzzzt, you are not The Authority.
In Win3.0 Window refered to just the window manager, which you ran from the OS called DOS.
by Win4.0, excuse me, Win 95, that name was adopted for the entire package, which still included DOS.
is "bzzzt" the sound your head makes when you try to think?
Thing is, people are calling this OS after the kernel, they can. Windows is named after it's windows manager. Many people call the OS they are using, "RedHat".
FSF could come out with a distro and call it GNU.
Further, the idea that GNU doesn't get credit is'nt really fair, how long can you use Linux without reading GNU this and GNU that?
It's a test of freedom, and the software is winning.
... yeah, and he also suggests installing a Monarchy in Irag.
Does anybody believe that the power-elite in the US are holders of the democratic ideal?
OK, experiment over, time for the post mortem to explain it's failure.
that's an interesting challenge, I might try to do that.
Your emphasis on change, and the design representing one's anticiplation of change is very interesting and more or less exactly how I design.
I find object oriented syntax convenient. Perhaps I can give you a more detailed explanation at some later date.
I realize it looked like I was boasting, but I was not, there is nothing particularly special about working in the game industry except to support my comments about where the game industry was in terms of code reuse.
It seemed boastful only because of your personal insecurity and --- what... your name really is Pyrrho? oh, sorry.
as far as getting you started, I've read some of your exposition on this.
I don't say that OO does it better, I think this book, Design Patterns, which happens to be a part of OOD, deals with the subject matter appropriates. It does not advocate an extreme, it advocates and understanding of the implications of different kinds of relationships.
And the reason why is exactly what you mention, understanding the implications of various kinds of relationships (aka coupling between components) is what's important, so that in each case you can decide what you need. Tight coupling is good in some cases, loose in others. And of course, there is gradiant from one extreme to the other.
On the other hand, I do think in terms of OOD, but not in the way you think of OOD. My OOD is much more pragmatic than the OOD that seems to have become a pet peeve of yours.
Believe me, I know what irks you in the sense that there certainly is a class of OOProgrammer that give OOD a bad name. That is, bad designs defended as good because they are object oriented.
astronomers need programmers.
there is paradigm bias... they understand data and functions as coupled, there is an OOD bias, but not bigotry.
Just as there is a bit of a C++ bias, but that's just a pragmatic factor.
The way Design Patterns handled coupling issues is very interesting. They don't say something like sickOOPs say, "decouple everything" "abstract everything", etc., but they tell you ways to decouple, and give insight on the benefits and costs of certain kinds of dependencies from a purely logical point of view.
Why not a C+SQL pattern books... well, there is room there for such a book. Problem: the "patterns" area of books is a sick wasteland, Design Patterns isn't just the shining star, it's practically the only decent example (though no doubt someone can point me to some exception, this is imho, ymmv).
(1) that doesn't change my opinion.
(2) I've seen many software projects.
by the way, the waterfall method is long since discredited, so if the logician doesn't have anything to do with the design of his code then he is at a bad company, a company that would make even interesting projects boring and pointless. Did you consider that?
you are supposed to separate the formatting/style... NOT DISCARD IT!
"Ok, I took the dirty diaper off the baby."
[looks around] "um, where's the baby"
"oh, what, I was supposed to keep the baby?"
The internet and shrink wrap software business mayhem have ruined the sensibilities of many programmers to the point they don't realize...
When people pay high $$$ for an "engineer", they WANT to be told how to do it right! And even cheap tech labor is expensive.
But after working for 23 year olds that just got 50G$ from some VC, they think everybody is an impetuous youth that wants the impossible, and if they can't have it, they would prefer a quick piece of crud to a tool that actually works!
When I read your sig, or the version that changes it to "slashdot" (don't know if that's you or not), I want to laugh, I get it... but everytime my compulsive gaurdian geek pops into view on my shoulder and says, "A million is no where near infinity... compared to infinity, a million is no more than 1."
but anyway, more on topic... how was the Grappa?
...where do I get the linux and solaris version?
All those people complaining about MS just didn't know about this little gem.
well, tablizer, having read your arguments about OOP your comment is not suprising.
But consider... Design Patterns is a logic book, it's about logical relations, and coupling, loose and tight, about references, linked, embeded, direct and indirect.
They are not templates. That is, you can impliment a pattern using SQL and C if you so desire! The implimentation is left for your (or some vender that wants to serve your needs), and that's the way it should be.
Actually Design Patterns (not talking about other pattern books, just Design Patterns) is probably the best book for you in particular as it doesn't say that OOD is the be all end all. In fact, when they give their example of a OO parser, they specifically say this is not the best way to build a parser, there are more efficient and standard structured approaches, and the pattern is presented to investigate the logic and functional relationships involved.
All software used to be this way.
In fact, games are no longer this way, or at least, it's not the profitable way.
Games need to use libraries, need to have engines that can be brought forward instead of thrown out.
In the old days the need to optimize for slow machines was a big pressure, you didn't have a general purpose 3D engine, you had something that produced a picture and underneath was really a bunch of special optimization tricks you didn't want to bring forward.
Now, the natural progression of reuse is entering the game industry. Look how well the Sims is doing, partially because they can release the Date Pack (or whatever) and improve their system instead of replace it.
and yes, I worked in the game industry (network focussed) for over 6 years.
want to impressed by versatility, read some of Michael Jackson's books, like "Problem Frames".
He's a much better systems analyst than he is musician. Or... it might be two different people (namespace collision).
Calling ReconRich,
The internet bubble has burst. Time to market over quality was a part of that bubble.
message over.
PS: to be serious, you do have a point, and time to market will always be important, but it's not the "time it takes me to have something I can put in the market" it will return to being "the time it takes to put together something that WORKS, and get it to market."
(1) I think there is more tedium than excitment protecting a village from bandits.
(2) All software projects can be interesting: if you are a logician. If you love logic. All these system are interesting to design and construct (less interesting, imho, to maintain and operate!)
Of course, I'm a bit full of it as I've always gravitate to parts of the industry that interest me, because it does help one motivate oneself, not so much for projects that are more logically interesting, but to stay motivated through political storms and other meatspace crud.
They've had to "uncouple" IE. Now they're being asked to uncouple Media Player.
What's next? Uncoupling the calculator? The start button? Command prompt?
Following this line of thinking ad absurdum, what exactly is Microsoft allowed to package with Windows? Sheesh!
Replaceable componentry is usually advisable based on engineering principle, but in this case legal systems are imposing a component based approach, attempting to protect their free markets.