What you want to look at for processing XML in a "lisp-like language" (Scheme) is SXML and its related packages (SXPath, etc.)
These are packages for manipulating and representing XML content in Scheme programs.
XML can be seen as a syntax tree, right? XSL and its friends are tree manipulation tools.
The same can be said for Lisp and Lisp-like languages, whose "program as data" philosophy (and 35 years of history) focuses on program evaluation as tree transformation and manipulation (through nested lists.)
SXML is a translation of XML from its heavy angle-bracket syntax to a Scheme sexpr (Scheme/Lisp's bracket expression syntax) syntax (and back to XML again.) It's extremely powerful.
It's worth looking at if you spend large amounts of time manipulating XML data.
When I was a kid in the early 80s and I read about "object orientation" and especially in reference to graphic user interfaces I really expected much more from them -- I remember my first encounters with Macintosh, Windows, Presentation Manager, etc. I was very disappointed. I was similarily disappointed years later when I finally got to play with the Smalltalk GUI that I had read about.
The only UI that impressed me was the Hypercard UI in combination with Hypertalk.
It has always seemed to me that there is such an incredible disjunction between the object-oriented infrastructure of some applications and the presentation they choose.
If things are truly object oriented, in the sense that objects are responding to messages and inheriting behaviours from each other, as a user of an application I ought to be able to send those messages. I also ought to be able to swap components out and visually replace them with other objects matching the same interface. I ought to be able to bring up a window and use a Smalltalk-style Inspector or Browser and visually see the relationship of objects in the application.
In an ideal world I ought also to be able to mouse over something, and bring up a _real_ "dialog box" and actually enter in messages for that object to invoke and see their responses. This seems similar to what the author of this slashdot submission was describing. (But to me it's less about "command line vs non-command line" as its about authors opening up the infrastructure of their apps for extension, scripting and inspection in general.)
But instead, GUI applications are really quite static, made out of prefab generic components. As an application author using an OO language I can't even use genuinely OO techniques with the GUI components: their implementation is frozen, you typically can't even really inherit from them (or if you can, you can't really change the visual implementation very much).
The only system that I've seen that defies this general trend is the Morphic graphics system in Squeak Smalltalk (but developed on Sun's Self project.) It's the only system I've seen that seems to focus on GUIs as environments (focused around objects) for exploration and customization rather than set, static applications where the focus is on "running the program" instead of on "using the objects."
Sometimes I feel we've made very little progress in the last 25 years:-(
It's not hard to replace the 720k drive with a standard modern high density floppy. There are quite a few text files on various Atari ST FTP sites that describe the mod.
Oh, and AFAIK formatting a high density floppy in a 720k drive is no problem if you're willing to deal with the smaller size.
I agree with your overall philosophy but not with some of your specific points.
You make the argument that Smalltalk and Lisp systems from the 70s were addressing the kinds of things that Microsoft and Apple have been dolling out in small doses over the years (inconsistently etc.) and I agree...
However, the deeper fact of the matter is Smalltalk wasn't and isn't a complete solution (and still aren't) and can't compete with, say, OS X, because it certainly doesn't provide:
* Component document models like OpenDoc was supposed to provide. Smalltalk-80 didn't have this, nor does Squeak.
* Further, they don't even really have tools like word processors, image editors, and other areas that normally fall in the "application" domain. I've seen simple web browsers and simple editors, but nothing on the order of functionality provided by most word processors. There is the argument that if you were to do these things properly object oriented they wouldn't look like Word anyways, but Smalltalk didn't/doesn't even do that.
* The same kinds of areas of work that are being heavily investigated in other languages are areas of heavy investigation and work in the platforms you mention, too: object distribution, published components and packages (Squeak is only now breaking out the single-image-single-system and allowing for modular ), persistence, etc.
* Performance of Smalltalk in the 80s was hardly acceptible on personal computers of the time. Apple and NeXT used the "compromised" (i.e. compiled, not interpreted, etc.) tech they did because they were running on 68K platforms with consumer hardware.
C++, Objective-C, Java, etc. are different answers to different questions. I don't necessarily agree with the questions asked, but they are solutions.
C++ is a great system-level programming language. It does a lot of things really well, though object-orientation isn't really one of them. Objective-C is a reasonable compromise for moving people from a C/Unix world to an OO world. Java is, well... an overarchitected and overmarkted approach to the same thing Smalltalk was attempting...
The problem is that the best system based on the principles established by PARC, etc. hasn't been built yet. Perhaps it could have if certain companies had displayed more guts and leadership -- but there were very distinct market forces acting against the progress of properly engineered, heavily researched, elegant technologies.
Actually, sorry, POUM (Partido Obrero de Unificación Marxista) were an organization that was aligned (somewhat tenuously) with the Fourth International and was composed of a variety of non-Stalinist-aligned communists, mostly Trotskyists council communists. They were aligned with the anarcho-syndicalist groups, definitely.
You know, your abusive tone wouldn't even be justified if you weren't totally and utterly wrong.
Firstly, only the United States and a few of its militarily armed-to-the-teeth, imperialist, and protectionist friends have done well economically under neo-liberalism -- and even they have had declining profitability rates since the growth periods of the 50s and 60s.
The vast majority of countries that have followed the recommended World Bank/IMF path of deregulation and privatization (i.e. capitalization) have actually suffered significantly in terms of overall spending power and standard of living.
Secondly, Eastern European countries weren't socialist, or communist. They were Stalinist bureaucracies ruled by a dictatorship of former (or sometimes not so former) communists with imperial ambitions.
What's so "socialist" about France, other than the fact that the ruling party for years called itself Socialist?
BTW, Marx cut his teeth on the critique of Prussian and Bonapartist (French) bureaucracy in the 1840s... I believe his favoured term for the bureacracy was "social parasites."
Take a look at graphs worldwide caloric intake counts for the last 50 years (hint, they drop significantly, especially in the last 20 years of neo-liberal deregulation and globalization) then tell me again about your beautiful system...
If you'd actually read either of these books with any sense of history or context, you'd see that both books (especially Animal Farm) are actually just fictive rewritings of Trotsky's "The Revolution Betrayed".
But, hey, I shouldn't talk over you -- you might need to learn who Trotsky is, first.
Reportedly Orwell just switched the last two numbers of the year of its writing. It wasn't meant to be a warning of the future, but a critique of the present (Stalinism, etc.)
Actually, I know someone (very well) who has the sad fortune of working there (3web/cybersurf.) They will be charging $10/month for their service, minus banner ads, and yanking the free service. They are in the process of doing this right now. And the really twisted thing is that there's talk that after yanking the free service and charging $10/month for the non-ads service, they may introduce ads again later.
Just another fucked up, loony "no business plan" company born in Alberta's wacked out "entrepreneurial" business climate. That company is so f*cked up, I give them maybe 6-8 months, if that, before they go down the tube completely. Put it this way, everyone in their Toronto office is trying to get out of there...
LambdaMOO! (a MUD with a single inheritance prototype based object oriented language... see http://www.moo.mud.org/ and telnet://lambda.moo.mud.org:8888/ for more details)
Squeak! (original Smalltalk-80 updated with multimedia extensions, a new prototype based UI called Morphic, ability to play Flash, Quicktime, etc.)
Either would teach important concepts of object orientation, prototyping, simple clean syntax, and have direct concrete results that younger people love (in LambdaMOO you can write methods on your "things" to make them do things in the room you're in, etc. and get immediate feedback)
I agree with the comment that they should do a prequel -- but they would have to try hard to maintain the original naivety of the film. It should go back to the early 80s style.. use retro-futurist elektro music (they should make the title track "Little Computer People" by Anthony Rother ("little computer people project")!!) See: show flynn starting up his arcade...
the originally movie was miserable except for its cuteness value.. it had bad acting, horrible plot, but the look of the whole thing. yummy simple look to the whole thing. and the wendy carlos soundtrack was amazing.
This is an excellent, well-thought out article, so all you nay-sayers can... I think you get the picture. I have been waiting for somebody out there to point out the SOCIAL CONTEXT of the "Y2K Problem" for months -- instead all I get is moronic techno-visionary bullshit about the millenium, or on the other side of the equasion, computer illiterate media morons blathering about issues they don't understand.
Let's be clear: a small percentage of the world's population is profiting at the expense of the majority. Until technology becomes a tool for aleviating that problem (which it could be, but isn't), we should be subjecting its industry to ruthless criticism. It is capitalism as a system that is the problem, and we need to use our combined knowledge and energy to make the next millenium one that belongs to the entire human race, rather than to a few select masters.
Slashdot Mirror
What you want to look at for processing XML in a "lisp-like language" (Scheme) is SXML and its related packages (SXPath, etc.)
These are packages for manipulating and representing XML content in Scheme programs.
XML can be seen as a syntax tree, right? XSL and its friends are tree manipulation tools.
The same can be said for Lisp and Lisp-like languages, whose "program as data" philosophy (and 35 years of history) focuses on program evaluation as tree transformation and manipulation (through nested lists.)
SXML is a translation of XML from its heavy angle-bracket syntax to a Scheme sexpr (Scheme/Lisp's bracket expression syntax) syntax (and back to XML again.) It's extremely powerful.
It's worth looking at if you spend large amounts of time manipulating XML data.
When I was a kid in the early 80s and I read about "object orientation" and especially in reference to graphic user interfaces I really expected much more from them -- I remember my first encounters with Macintosh, Windows, Presentation Manager, etc. I was very disappointed. I was similarily disappointed years later when I finally got to play with the Smalltalk GUI that I had read about.
:-(
The only UI that impressed me was the Hypercard UI in combination with Hypertalk.
It has always seemed to me that there is such an incredible disjunction between the object-oriented infrastructure of some applications and the presentation they choose.
If things are truly object oriented, in the sense that objects are responding to messages and inheriting behaviours from each other, as a user of an application I ought to be able to send those messages. I also ought to be able to swap components out and visually replace them with other objects matching the same interface. I ought to be able to bring up a window and use a Smalltalk-style Inspector or Browser and visually see the relationship of objects in the application.
In an ideal world I ought also to be able to mouse over something, and bring up a _real_ "dialog box" and actually enter in messages for that object to invoke and see their responses. This seems similar to what the author of this slashdot submission was describing. (But to me it's less about "command line vs non-command line" as its about authors opening up the infrastructure of their apps for extension, scripting and inspection in general.)
But instead, GUI applications are really quite static, made out of prefab generic components. As an application author using an OO language I can't even use genuinely OO techniques with the GUI components: their implementation is frozen, you typically can't even really inherit from them (or if you can, you can't really change the visual implementation very much).
The only system that I've seen that defies this general trend is the Morphic graphics system in Squeak Smalltalk (but developed on Sun's Self project.) It's the only system I've seen that seems to focus on GUIs as environments (focused around objects) for exploration and customization rather than set, static applications where the focus is on "running the program" instead of on "using the objects."
Sometimes I feel we've made very little progress in the last 25 years
It's not hard to replace the 720k drive with a standard modern high density floppy. There are quite a few text files on various Atari ST FTP sites that describe the mod.
Oh, and AFAIK formatting a high density floppy in a 720k drive is no problem if you're willing to deal with the smaller size.
I agree with your overall philosophy but not with some of your specific points.
You make the argument that Smalltalk and Lisp systems from the 70s were addressing the kinds of things that Microsoft and Apple have been dolling out in small doses over the years (inconsistently etc.) and I agree...
However, the deeper fact of the matter is Smalltalk wasn't and isn't a complete solution (and still aren't) and can't compete with, say, OS X, because it certainly doesn't provide:
* Component document models like OpenDoc was supposed to provide. Smalltalk-80 didn't have this, nor does Squeak.
* Further, they don't even really have tools like word processors, image editors, and other areas that normally fall in the "application" domain. I've seen simple web browsers and simple editors, but nothing on the order of functionality provided by most word processors. There is the argument that if you were to do these things properly object oriented they wouldn't look like Word anyways, but Smalltalk didn't/doesn't even do that.
* The same kinds of areas of work that are being heavily investigated in other languages are areas of heavy investigation and work in the platforms you mention, too: object distribution, published components and packages (Squeak is only now breaking out the single-image-single-system and allowing for modular ), persistence, etc.
* Performance of Smalltalk in the 80s was hardly acceptible on personal computers of the time. Apple and NeXT used the "compromised" (i.e. compiled, not interpreted, etc.) tech they did because they were running on 68K platforms with consumer hardware.
C++, Objective-C, Java, etc. are different answers to different questions. I don't necessarily agree with the questions asked, but they are solutions.
C++ is a great system-level programming language. It does a lot of things really well, though object-orientation isn't really one of them. Objective-C is a reasonable compromise for moving people from a C/Unix world to an OO world. Java is, well... an overarchitected and overmarkted approach to the same thing Smalltalk was attempting...
The problem is that the best system based on the principles established by PARC, etc. hasn't been built yet. Perhaps it could have if certain companies had displayed more guts and leadership -- but there were very distinct market forces acting against the progress of properly engineered, heavily researched, elegant technologies.
*Shrug*, you're a mindless cynic, instead.
Actually, sorry, POUM (Partido Obrero de Unificación Marxista) were an organization that was aligned (somewhat tenuously) with the Fourth International and was composed of a variety of non-Stalinist-aligned communists, mostly Trotskyists council communists. They were aligned with the anarcho-syndicalist groups, definitely.
You know, your abusive tone wouldn't even be justified if you weren't totally and utterly wrong. Firstly, only the United States and a few of its militarily armed-to-the-teeth, imperialist, and protectionist friends have done well economically under neo-liberalism -- and even they have had declining profitability rates since the growth periods of the 50s and 60s. The vast majority of countries that have followed the recommended World Bank/IMF path of deregulation and privatization (i.e. capitalization) have actually suffered significantly in terms of overall spending power and standard of living. Secondly, Eastern European countries weren't socialist, or communist. They were Stalinist bureaucracies ruled by a dictatorship of former (or sometimes not so former) communists with imperial ambitions.
What's so "socialist" about France, other than the fact that the ruling party for years called itself Socialist?
BTW, Marx cut his teeth on the critique of Prussian and Bonapartist (French) bureaucracy in the 1840s... I believe his favoured term for the bureacracy was "social parasites."
Umm, keep dreaming.
Take a look at graphs worldwide caloric intake counts for the last 50 years (hint, they drop significantly, especially in the last 20 years of neo-liberal deregulation and globalization) then tell me again about your beautiful system...
I'd be disappointed if Orwell didn't have a critique of Trotsky himself (god knows he needs one).
Doesn't change the fact that his basic sympathies remained with the "animals" -- just not with the pigs (or the farmers.)
Nonsene.
If you'd actually read either of these books with any sense of history or context, you'd see that both books (especially Animal Farm) are actually just fictive rewritings of Trotsky's "The Revolution Betrayed".
But, hey, I shouldn't talk over you -- you might need to learn who Trotsky is, first.
Reportedly Orwell just switched the last two numbers of the year of its writing. It wasn't meant to be a warning of the future, but a critique of the present (Stalinism, etc.)
That's funny, because he was a socialist. He fought in the Spanish Civil War in a Trotskyist militia. Get your facts straight, righty-boy.
Just another fucked up, loony "no business plan" company born in Alberta's wacked out "entrepreneurial" business climate. That company is so f*cked up, I give them maybe 6-8 months, if that, before they go down the tube completely. Put it this way, everyone in their Toronto office is trying to get out of there...
Ryan
Nothing like a blind right winger with an agenda. Go hang out with your buddy Stockwell.
I would suggest either of the following
... see http://www.moo.mud.org/ and telnet://lambda.moo.mud.org:8888/ for more details)
LambdaMOO! (a MUD with a single inheritance prototype based object oriented language
Squeak! (original Smalltalk-80 updated with multimedia extensions, a new prototype based UI called Morphic, ability to play Flash, Quicktime, etc.)
Either would teach important concepts of object orientation, prototyping, simple clean syntax, and have direct concrete results that younger people love (in LambdaMOO you can write methods on your "things" to make them do things in the room you're in, etc. and get immediate feedback)
ryan
oops: check out http://members.xoom.com/_XOOM/psi49net/sou/lcp1_is dn_txt.ram
I agree with the comment that they should do a prequel -- but they would have to try hard to maintain the original naivety of the film. It should go back to the early 80s style.. use retro-futurist elektro music (they should make the title track "Little Computer People" by Anthony Rother ("little computer people project")!!) See: show flynn starting up his arcade...
.. it had bad acting, horrible plot, but the look of the whole thing. yummy simple look to the whole thing. and the wendy carlos soundtrack was amazing.
the originally movie was miserable except for its cuteness value
ahh memories of grade 2
This is an excellent, well-thought out article, so all you nay-sayers can... I think you get the picture. I have been waiting for somebody out there to point out the SOCIAL CONTEXT of the "Y2K Problem" for months -- instead all I get is moronic techno-visionary bullshit about the millenium, or on the other side of the equasion, computer illiterate media morons blathering about issues they don't understand.
Let's be clear: a small percentage of the world's population is profiting at the expense of the majority. Until technology becomes a tool for aleviating that problem (which it could be, but isn't), we should be subjecting its industry to ruthless criticism. It is capitalism as a system that is the problem, and we need to use our combined knowledge and energy to make the next millenium one that belongs to the entire human race, rather than to a few select masters.
ryan