Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.forth > #36512 > unrolled thread

Is 2015 the year of FORTH ?

Back to article view | Back to comp.lang.forth

Contents

  Is 2015 the year of FORTH ? visualforth@rocketmail.com - 2015-02-12 12:13 -0800
    Is 2015 the year of FORTH ? Mark Wills <markwills1970@gmail.com> - 2015-02-12 12:56 -0800
      Re: Is 2015 the year of FORTH ? empty-buffers <empty.buffersYOU-DONT-WANT-THIS@gmail.NEITHER-THIS.com> - 2015-02-12 21:32 +0000
        Re: Is 2015 the year of FORTH ? rickman <gnuarm@gmail.com> - 2015-02-13 13:08 -0500
          Re: Is 2015 the year of FORTH ? Bernd Paysan <bernd.paysan@gmx.de> - 2015-02-13 19:39 +0100
            Re: Is 2015 the year of FORTH ? rickman <gnuarm@gmail.com> - 2015-02-13 14:14 -0500
              Re: Is 2015 the year of FORTH ? Bernd Paysan <bernd.paysan@gmx.de> - 2015-02-13 21:28 +0100
                Re: Is 2015 the year of FORTH ? Jason Damisch <jasondamisch@yahoo.com> - 2015-02-13 14:22 -0800
                  Re: Is 2015 the year of FORTH ? Bernd Paysan <bernd.paysan@gmx.de> - 2015-02-14 01:34 +0100
                Re: Is 2015 the year of FORTH ? hughaguilar96@gmail.com - 2015-02-13 22:42 -0800
                HTML vs. LaTeX (was: Is 2015 the year of FORTH ?) anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2015-02-14 10:21 +0000
                  Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?) Bernd Paysan <bernd.paysan@gmx.de> - 2015-02-14 19:37 +0100
                    Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?) visualforth@rocketmail.com - 2015-02-14 11:13 -0800
                      Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?) Bernd Paysan <bernd.paysan@gmx.de> - 2015-02-14 21:07 +0100
                        Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?) Jason Damisch <jasondamisch@yahoo.com> - 2015-02-14 12:50 -0800
                          Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?) Bernd Paysan <bernd.paysan@gmx.de> - 2015-02-14 22:32 +0100
                            Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?) Jason Damisch <jasondamisch@yahoo.com> - 2015-02-14 15:10 -0800
                    Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?) anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2015-02-15 10:34 +0000
                      Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?) Bernd Paysan <bernd.paysan@gmx.de> - 2015-02-15 22:15 +0100
                        Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?) anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2015-02-16 12:45 +0000
                      Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?) krishna.myneni@ccreweb.org - 2015-02-15 13:55 -0800
                        Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?) anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2015-02-16 13:52 +0000
                          Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?) Bernd Paysan <bernd.paysan@gmx.de> - 2015-02-16 15:47 +0100
                      Re: HTML vs. LaTeX Paul Rubin <no.email@nospam.invalid> - 2015-02-16 09:05 -0800
                        Re: HTML vs. LaTeX anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2015-02-17 12:21 +0000
                          Re: HTML vs. LaTeX Paul Rubin <no.email@nospam.invalid> - 2015-02-20 13:23 -0800
                            Re: HTML vs. LaTeX Bernd Paysan <bernd.paysan@gmx.de> - 2015-02-20 22:33 +0100
                          Re: HTML vs. LaTeX anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2015-02-21 15:14 +0000
                            Re: HTML vs. LaTeX Bernd Paysan <bernd.paysan@gmx.de> - 2015-02-21 20:23 +0100
                            Re: HTML vs. LaTeX Paul Rubin <no.email@nospam.invalid> - 2015-02-21 11:33 -0800
                Re: Is 2015 the year of FORTH ? krishna.myneni@ccreweb.org - 2015-02-14 10:14 -0800
                  Re: Is 2015 the year of FORTH ? rickman <gnuarm@gmail.com> - 2015-02-22 01:57 -0500
                    Re: Is 2015 the year of FORTH ? krishna.myneni@ccreweb.org - 2015-02-22 18:35 -0800
                      Re: Is 2015 the year of FORTH ? rickman <gnuarm@gmail.com> - 2015-02-23 01:58 -0500
                        Re: Is 2015 the year of FORTH ? krishna.myneni@ccreweb.org - 2015-02-23 04:59 -0800
                          Re: Is 2015 the year of FORTH ? rickman <gnuarm@gmail.com> - 2015-02-23 19:18 -0500
                            Re: Is 2015 the year of FORTH ? krishna.myneni@ccreweb.org - 2015-02-23 18:09 -0800
                            Re: Is 2015 the year of FORTH ? Mark Wills <markwills1970@gmail.com> - 2015-02-24 01:04 -0800
                              Re: Is 2015 the year of FORTH ? krishna.myneni@ccreweb.org - 2015-02-24 02:51 -0800
                                Re: Is 2015 the year of FORTH ? Jason Damisch <jasondamisch@yahoo.com> - 2015-02-24 10:23 -0800
                      Re: Is 2015 the year of FORTH ? Bernd Paysan <bernd.paysan@gmx.de> - 2015-02-23 21:19 +0100
                        Re: Is 2015 the year of FORTH ? krishna.myneni@ccreweb.org - 2015-02-23 17:58 -0800
                          Re: Is 2015 the year of FORTH ? Mark Wills <markwills1970@gmail.com> - 2015-02-24 01:07 -0800
                            Re: Is 2015 the year of FORTH ? krishna.myneni@ccreweb.org - 2015-02-24 02:19 -0800
                Re: Is 2015 the year of FORTH ? Paul E Bennett <Paul_E.Bennett@topmail.co.uk> - 2015-02-15 12:18 +0000
          Re: Is 2015 the year of FORTH ? empty-buffers <empty.buffersYOU-DONT-WANT-THIS@gmail.NEITHER-THIS.com> - 2015-02-13 18:55 +0000
            Re: Is 2015 the year of FORTH ? Melzzzzz <mel@zzzzz.com> - 2015-02-13 20:27 +0100
          Re: Is 2015 the year of FORTH ? "Elizabeth D. Rather" <erather@forth.com> - 2015-02-13 09:53 -1000
    Re: Is 2015 the year of FORTH ? "WJ" <w_a_x_man@yahoo.com> - 2015-02-13 14:18 +0000
      Re: Is 2015 the year of FORTH ? visualforth@rocketmail.com - 2015-02-13 11:03 -0800
      Re: Is 2015 the year of FORTH ? hughaguilar96@gmail.com - 2015-02-13 22:46 -0800
      Re: Is 2015 the year of FORTH ? Mark Wills <markwills1970@gmail.com> - 2015-02-14 02:35 -0800
        Re: Is 2015 the year of FORTH ? Coos Haak <chforth@hccnet.nl> - 2015-02-14 18:47 +0100
        Re: Is 2015 the year of FORTH ? hughaguilar96@gmail.com - 2015-02-14 18:56 -0800
    Re: Is 2015 the year of FORTH ? Ilya Tarasov <ilya74.tarasov@gmail.com> - 2015-02-15 17:10 -0800
      Re: Is 2015 the year of FORTH ? Mark Wills <markwills1970@gmail.com> - 2015-02-16 01:20 -0800
        Re: Is 2015 the year of FORTH ? Lars Brinkhoff <lars.spam@nocrew.org> - 2015-02-16 12:35 +0100
        Re: Is 2015 the year of FORTH ? "Elizabeth D. Rather" <erather@forth.com> - 2015-02-16 08:12 -1000
        Re: Is 2015 the year of FORTH ? Ilya Tarasov <ilya74.tarasov@gmail.com> - 2015-02-16 15:17 -0800

Page 1 of 3  [1] 2 3  Next page →

#36512 — Is 2015 the year of FORTH ?

Altia has a new resource available today -- an eBook on the Top Embedded Trends for 2015. Their in-house industry experts have complied five trends to look out for in the coming year.

Trend #4 (sic!) is:

"Minimalism"
"As technology steadily advances, designers and developers will seek ways to counter that complexity with simplicity. The sum of a user's microinteractions quickly defines that individual device's brand and product perception. A focus on clean, uncomplicated design will enable users to perform complex tasks on their devices with unprecedented ease."

That's really promising!
Sounds like Forth!
Let's make it happen!

Source:
http://connect.altia.com/hs-fs/hub/260438/file-2471049867-pdf/5_Embedded_Trends_ebook_v4.pdf

[toc] | [next] | [standalone]

#36517

Meh. I'll believe it when I see it.  You could apply the exact same reasoning to any year. Nothing special whatsoever about 2015.

[toc] | [prev] | [next] | [standalone]

#36520

In comp.lang.forth, Mark Wills wrote:

> Meh. I'll believe it when I see it.  You could apply the exact same
> reasoning to any year.  Nothing special whatsoever about 2015.

I'm afraid, we won't see any breakthru. From what I understand, the
popularity of C(-derivatives) mostly came from the fact, that the industry
wants easily replaceable - even, if at most average - programmers; on one
hand they use libraries written by other replaceable programmers, on the
other hand the "patronizing" compiler always fixes their most obvious errors.

Although software created by legions of such programmers still grows with
fat (we never notice the opposite; still the need for more RAM and more
powerful CPUs), the consumers can use that bloaty software exactly because
hardware becomes each year more powerful. "We can afford this".

If I'm correct, in case of Forth we could see exactly the opposite: much
less resource-hungry programs, but its authors - in case of "need" - couldn't
be as easily sacked and replaced with others. I believe, no management
likes this.
-- 
"Yes, this is the bad end you've come to, and two and two is - as it always
was..." - began Trurl, but just then the machine made a faint, barely audible
croaking noise and said, for the last time: "SEVEN".

[toc] | [prev] | [next] | [standalone]

#36544

On 2/12/2015 4:32 PM, empty-buffers wrote:
> In comp.lang.forth, Mark Wills wrote:
>
>> Meh. I'll believe it when I see it.  You could apply the exact same
>> reasoning to any year.  Nothing special whatsoever about 2015.
>
> I'm afraid, we won't see any breakthru. From what I understand, the
> popularity of C(-derivatives) mostly came from the fact, that the industry
> wants easily replaceable - even, if at most average - programmers; on one
> hand they use libraries written by other replaceable programmers, on the
> other hand the "patronizing" compiler always fixes their most obvious errors.
>
> Although software created by legions of such programmers still grows with
> fat (we never notice the opposite; still the need for more RAM and more
> powerful CPUs), the consumers can use that bloaty software exactly because
> hardware becomes each year more powerful. "We can afford this".
>
> If I'm correct, in case of Forth we could see exactly the opposite: much
> less resource-hungry programs, but its authors - in case of "need" - couldn't
> be as easily sacked and replaced with others. I believe, no management
> likes this.

I think this is a bit of a negative image.  It's not so much that they 
want to sack people as they want to be able to hire good people.  If you 
have many fewer to choose from you are not as able to be selective.

As to the whole bloatware thing, I think that is more a matter of being 
able to do things you couldn't before because now you have the resources 
to enable new methods.  For example it was explained to me that one of 
the memory grabbing features of browsers is the large memory buffers 
they allocate just to hold complete images of your pages.  I guess this 
speeds up context switches and such.  Without many MB of RAM available 
that would not be possible and many things would happen more slowly.

Bloatware is one of those strawmen that we create and attack to prove a 
point without actually showing any facts.  I seriously doubt that a 
browser with similar capabilities to other browsers would use any less 
memory if written in Forth.

-- 

Rick

[toc] | [prev] | [next] | [standalone]

#36545

rickman wrote:
> Bloatware is one of those strawmen that we create and attack to prove a
> point without actually showing any facts.  I seriously doubt that a
> browser with similar capabilities to other browsers would use any less
> memory if written in Forth.

The accidential complexity of such programs is *far bigger* than the 
inherent complexity. E.g. if you compare the typesetting capabilities of a 
modern browser with LaTeX, you see that LaTeX can do a lot more in a lot 
less memory.  Why?  Because the core of TeX was written by one clever guy 
who knows his O() in terms of speed and size - he knows how to chose 
algorithms and datastructures wisely.  He even wrote some books about the 
art of computer programming.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://bernd-paysan.de/
net2o ID: kQusJzA;7*?t=uy@X}1GWr!+0qqp_Cn176t4(dQ*

[toc] | [prev] | [next] | [standalone]

#36548

On 2/13/2015 1:39 PM, Bernd Paysan wrote:
> rickman wrote:
>> Bloatware is one of those strawmen that we create and attack to prove a
>> point without actually showing any facts.  I seriously doubt that a
>> browser with similar capabilities to other browsers would use any less
>> memory if written in Forth.
>
> The accidential complexity of such programs is *far bigger* than the
> inherent complexity. E.g. if you compare the typesetting capabilities of a
> modern browser with LaTeX, you see that LaTeX can do a lot more in a lot
> less memory.  Why?  Because the core of TeX was written by one clever guy
> who knows his O() in terms of speed and size - he knows how to chose
> algorithms and datastructures wisely.  He even wrote some books about the
> art of computer programming.

You don't make any sense.  Latex is not a browser.  It's not even 
remotely like a browser.  The only thing it has in common is that it can 
create an image on your screen of a given page design.  I think there is 
a lot more to browsers than what Latex does.  Otherwise why would they 
not just use the Latex language for browsers?

-- 

Rick

[toc] | [prev] | [next] | [standalone]

#36552

rickman wrote:

> On 2/13/2015 1:39 PM, Bernd Paysan wrote:
>> rickman wrote:
>>> Bloatware is one of those strawmen that we create and attack to prove a
>>> point without actually showing any facts.  I seriously doubt that a
>>> browser with similar capabilities to other browsers would use any less
>>> memory if written in Forth.
>>
>> The accidential complexity of such programs is *far bigger* than the
>> inherent complexity. E.g. if you compare the typesetting capabilities of
>> a modern browser with LaTeX, you see that LaTeX can do a lot more in a
>> lot
>> less memory.  Why?  Because the core of TeX was written by one clever guy
>> who knows his O() in terms of speed and size - he knows how to chose
>> algorithms and datastructures wisely.  He even wrote some books about the
>> art of computer programming.
> 
> You don't make any sense.  Latex is not a browser.  It's not even
> remotely like a browser.  The only thing it has in common is that it can
> create an image on your screen of a given page design.  I think there is
> a lot more to browsers than what Latex does.  Otherwise why would they
> not just use the Latex language for browsers?

One part of what the browser does is typesetting a structured text document.  
Instead of using LaTeX, they use HTML+CSS, but it is typesetting a 
structured text document - into a window instead of on paper, but that 
difference is small (you just don't need the page breaking algorithm).  And 
to your question: They reinvented half of what LaTeX can in a painstakenly 
process which took 20 years.  There's now even MathML in the browser to 
typeset formulas.  What they yet didn't invent was a real typesetting 
language, which would allow to load extensions to the typesetting into the 
browser.  So while LaTeX can have tons of styles on your harddisk, which you 
never need, the browser needs to load all his capabilities at start.

You better ask Tim Berner-Lee why he invented HTML, and didn't just put 
LaTeX into his browser; the latter would have made a lot of sense, since all 
the people at CERN usually write their stuff in LaTeX anyways.  But it would 
have been more work upfront.

Probably Tim thought, using HTML would make his job easier, because he did 
just implement a tiny subset of LaTeX.  At that time, it was simpler, and he 
could even integrate an interactive web page editor, because it was so 
simple.  And then, they had to expand this tiny thing to a full-blown 
typsetting engine with styles, with math, and a lot of other stuff.  And by 
doing that as an organic growth rather than with a structured approach, they 
created a lot of bloat.

IMHO that's because they did the browser upside down.  They started with the 
highest level, the structured hypertext.  And then, they slowly integrated 
the lower levels, which makes the whole thing flexible enough for all the 
diverse tasks a browser has to accomplish.  They added a style language - 
CSS.  They added a programming language, JavaScript.  They now even added a 
subset of OpenGL 1.0 in the form of WebGL, a rendering infrastructure.  Not 
the good one, which starts at OpenGL 2, and comes with a shader language, 
just the very basic thing.

And of course, by doing it the wrong way round, it all gets extremely 
complex.

I'm going to implement a browser for my net2o protocol.  I won't start on 
the top, the starting point, the "primitives" to build a browser are things 
like OpenGL, FreeType, and image decoder (SOIL), video decoding, and of 
course, a low-level, 100% general purpose programming language (Forth in a 
Sandbox).  Sooner or later, you'll get that anyways, there's asm.js, a 
portable assembler in JavaScript.  And the rest towards higher levels of 
abstraction are just programs you can load and execute in that sandbox.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://bernd-paysan.de/
net2o ID: kQusJzA;7*?t=uy@X}1GWr!+0qqp_Cn176t4(dQ*

[toc] | [prev] | [next] | [standalone]

#36555

On Friday, February 13, 2015 at 12:28:39 PM UTC-8, Bernd Paysan wrote:

> You better ask Tim Berner-Lee why he invented HTML, and didn't just put 
> LaTeX into his browser; the latter would have made a lot of sense, since all 
> the people at CERN usually write their stuff in LaTeX anyways.  But it would 
> have been more work upfront.

This is interesting because I do websites for a living, not programming in Java or Forth or anything like that.  Forth is more of a never-ending side project for me at the moment.  I've programmed some HTML code through a .dll using Forth.  I also wonder why the current setup with HTML5, CSS, JS, AJAX etc is so bloated and sometimes difficult to work with.

> And by 
> doing that as an organic growth rather than with a structured approach, they 
> created a lot of bloat.

I would use maybe the opposite analogy, but that is the way I think, its my own framework.  That is to say the bottom up approach is organic and the top down approach with a great deal of disconnect between what the creators of a standard have specified, and how the various systems are used by completely different people who have no say in how the systems are implemented is inorganic.

> IMHO that's because they did the browser upside down.  They started with the 
> highest level, the structured hypertext.  And then, they slowly integrated 
> the lower levels,

This is the mistake?

> which makes the whole thing flexible enough for all the 
> diverse tasks a browser has to accomplish.  They added a style language - 
> CSS.  They added a programming language, JavaScript.  They now even added a 
> subset of OpenGL 1.0 in the form of WebGL, a rendering infrastructure.  Not 
> the good one, which starts at OpenGL 2, and comes with a shader language, 
> just the very basic thing.

It is flexible because everybody must use it.  The way the world works is
that there is no real alternatives.  Everybody must use it so eventually it
will include everything, but not in an organic way.
 
> And of course, by doing it the wrong way round, it all gets extremely 
> complex.

Yes, thank you for the insight.
 
> I'm going to implement a browser for my net2o protocol.  I won't start on 
> the top, the starting point, the "primitives" to build a browser are things 
> like OpenGL, FreeType, and image decoder (SOIL), video decoding, and of 
> course, a low-level, 100% general purpose programming language (Forth in a 
> Sandbox).  Sooner or later, you'll get that anyways, there's asm.js, a 
> portable assembler in JavaScript.  And the rest towards higher levels of 
> abstraction are just programs you can load and execute in that sandbox.

I have to admit to being a little jealous of you. I am nowhere near as
advanced as you, being able to create your own protocol.  I also don't
know how to implement already specified protocols, such that Greg Bailey
has done.

I wonder if I will be able to find the time, energy, and brainpower to
investigate your setup?
 
> -- 
> Bernd Paysan
> "If you want it done right, you have to do it yourself"
> http://bernd-paysan.de/
> net2o ID: kQusJzA;7*?t=uy@X}1GWr!+0qqp_Cn176t4(dQ*

Bernd!

the URL leading at   net2o   takes you to a 404!

Jason

[toc] | [prev] | [next] | [standalone]

#36556

Jason Damisch wrote:
> I would use maybe the opposite analogy, but that is the way I think, its
> my own framework.  That is to say the bottom up approach is organic and
> the top down approach with a great deal of disconnect between what the
> creators of a standard have specified, and how the various systems are
> used by completely different people who have no say in how the systems are
> implemented is inorganic.

Yes.

>> IMHO that's because they did the browser upside down.  They started with
>> the highest level, the structured hypertext.  And then, they slowly
>> integrated the lower levels,
> 
> This is the mistake?

You described it above: instead of having organic growth, you have a 
disconnected top-down approach.  To get a reasonable result, you need a 
solid foundation.  Some primitives you can work with.  A natural form of 
extensibility you can build on.

> It is flexible because everybody must use it.  The way the world works is
> that there is no real alternatives.  Everybody must use it so eventually
> it will include everything, but not in an organic way.

The biggest mistake is IMHO that you can't define the extensions yourself.  
You can do some bits here and there with JavaScript, but it's merely inside 
the available framework.

> I have to admit to being a little jealous of you. I am nowhere near as
> advanced as you, being able to create your own protocol.  I also don't
> know how to implement already specified protocols, such that Greg Bailey
> has done.
> 
> I wonder if I will be able to find the time, energy, and brainpower to
> investigate your setup?
>  
>> --
>> Bernd Paysan
>> "If you want it done right, you have to do it yourself"
>> http://bernd-paysan.de/
>> net2o ID: kQusJzA;7*?t=uy@X}1GWr!+0qqp_Cn176t4(dQ*
> 
> Bernd!
> 
> the URL leading at   net2o   takes you to a 404!

That's not an URL, that's a public key (in base85).  It doesn't work outside 
the net2o protocol.

Or do you mean the URL on net2o.de leading to the fossil wiki?  Yes, 
apparently the Fossil authors have removed the shortcut paths for the wiki 
:-(.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://bernd-paysan.de/
net2o ID: kQusJzA;7*?t=uy@X}1GWr!+0qqp_Cn176t4(dQ*

[toc] | [prev] | [next] | [standalone]

#36561

On Friday, February 13, 2015 at 1:28:39 PM UTC-7, Bernd Paysan wrote:
> You better ask Tim Berner-Lee why he invented HTML, and didn't just put 
> LaTeX into his browser; the latter would have made a lot of sense, since all 
> the people at CERN usually write their stuff in LaTeX anyways.  But it would 
> have been more work upfront.

A lot of people have wondered about why HTML was gratuitously incompatible with LaTeX --- just think of all the time that have been spent over the years in converting between LaTeX and HTML (there are automated conversion programs, but they don't work very well) --- just think of how much useful work could have been done with that time instead.
 
> Probably Tim thought, using HTML would make his job easier, because he did 
> just implement a tiny subset of LaTeX.  At that time, it was simpler, and he 
> could even integrate an interactive web page editor, because it was so 
> simple.  And then, they had to expand this tiny thing to a full-blown 
> typsetting engine with styles, with math, and a lot of other stuff.  And by 
> doing that as an organic growth rather than with a structured approach, they 
> created a lot of bloat.
>
> IMHO that's because they did the browser upside down.  They started with the 
> highest level, the structured hypertext.  And then, they slowly integrated 
> the lower levels, which makes the whole thing flexible enough for all the 
> diverse tasks a browser has to accomplish.  They added a style language - 
> CSS.  They added a programming language, JavaScript.  They now even added a 
> subset of OpenGL 1.0 in the form of WebGL, a rendering infrastructure.  Not 
> the good one, which starts at OpenGL 2, and comes with a shader language, 
> just the very basic thing.
> 
> And of course, by doing it the wrong way round, it all gets extremely 
> complex.

True enough --- it is always better to start at the bottom and work up --- top-down has never been successful.

[toc] | [prev] | [next] | [standalone]

#36564 — HTML vs. LaTeX (was: Is 2015 the year of FORTH ?)

Bernd Paysan <bernd.paysan@gmx.de> writes:
>You better ask Tim Berner-Lee why he invented HTML, and didn't just put 
>LaTeX into his browser; the latter would have made a lot of sense, since all 
>the people at CERN usually write their stuff in LaTeX anyways.  But it would 
>have been more work upfront.

I think it would have been less work; the browser would have been just
wget (new, because it did not exist at the time. but simple to write),
latex and xdvi, with extensions for hyperlinks.  The problem would
have been that on the hardware of 1990 , it would have been too slow.
Our LaTeX benchmark takes 93s on a 486/66
<http://www.complang.tuwien.ac.at/franz/latex-bench>.  The best
hardware in 1990 was something like a 33MHz 486 or R3000, and
Berners-Lee had to target the majority of existing hardware, e.g.,
something like a 16MHz 386.  Ok, a web page would not take as much
time as our LaTeX benchmark, but still too much.

>Probably Tim thought, using HTML would make his job easier, because he did 
>just implement a tiny subset of LaTeX.  At that time, it was simpler, and he 
>could even integrate an interactive web page editor, because it was so 
>simple.  And then, they had to expand this tiny thing to a full-blown 
>typsetting engine with styles, with math, and a lot of other stuff.

Did they have to?  HTML 2.0 is fine for most of my uses, and 4.0
covers the rest (and provides lots of unneeded fluff).  I don't need
styles, and I am pretty sure they were not needed for what Berners-Lee
had in mind, either.  So he took the Forth approach: Created something
new tailored for his needs instead of catering for future expansion.

Why he based HTML on SGML instead of on LaTeX or the kind of markup
used in Usenet postings is a good question.  HTML is worse than either
for writing stuff manually, and that has led to stuff like WYSIWYG
editors for HTML, which were pretty bad.  OTOH, TeX as a language is
also badly designed, so I can understand that he did not want to go
there.

Anyway, the Chuck Moore approach of throwing everything away upon
encountering new requirements and designing the new thing from scratch
was not taken for HTML.  Instead, the tried and proven methods of the
software industry were used, and they result in bloat.  The web is 25
years old, and has expanded in usage far beyond the original vision,
so it's no wonder that browsers are bloated (the most popular ones are
based on Mosaic from 1993, so even the software itself has seen
"organic growth" for 22 years).  And Opera also has a 20-year history.

And while users may moan about bloated software, they still buy it,
and let the slim competition die from lack of revenue.

- anton
-- 
M. Anton Ertl  http://www.complang.tuwien.ac.at/anton/home.html
comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
     New standard: http://www.forth200x.org/forth200x.html
   EuroForth 2014: http://www.euroforth.org/ef14/

[toc] | [prev] | [next] | [standalone]

#36570 — Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?)

Anton Ertl wrote:

> Bernd Paysan <bernd.paysan@gmx.de> writes:
>>You better ask Tim Berner-Lee why he invented HTML, and didn't just put
>>LaTeX into his browser; the latter would have made a lot of sense, since
>>all
>>the people at CERN usually write their stuff in LaTeX anyways.  But it
>>would have been more work upfront.
> 
> I think it would have been less work; the browser would have been just
> wget (new, because it did not exist at the time. but simple to write),
> latex and xdvi, with extensions for hyperlinks.  The problem would
> have been that on the hardware of 1990 , it would have been too slow.
> Our LaTeX benchmark takes 93s on a 486/66
> <http://www.complang.tuwien.ac.at/franz/latex-bench>.  The best
> hardware in 1990 was something like a 33MHz 486 or R3000, and
> Berners-Lee had to target the majority of existing hardware, e.g.,
> something like a 16MHz 386.  Ok, a web page would not take as much
> time as our LaTeX benchmark, but still too much.

Berner-Lee only wrote for the NeXT cubes they had available.  Your benchmark 
runs on a 140k byte LaTeX document, rendering 33 pages.  The typical web 
page of that time was one page.

Furthermore, using LaTeX, you could have prerendered all pages, and just 
transmitted the dvis.  xdvi really did perform very well, and was only 
slightly larger than the original LaTeX (212kB for the 140kB of your 
benchmark).  xdvi however was very slow on embedded images, since these 
images were embedded as PostScript.  But that could have been solved easily, 
just add another \special for gif and jpeg.

> Did they have to?  HTML 2.0 is fine for most of my uses, and 4.0
> covers the rest (and provides lots of unneeded fluff).  I don't need
> styles, and I am pretty sure they were not needed for what Berners-Lee
> had in mind, either.  So he took the Forth approach: Created something
> new tailored for his needs instead of catering for future expansion.

The original idea of Berner-Lee was to improve communication between 
scientists at CERN.  The sort of people who already write a lot of LaTeX.  
Maybe it would have been a bad idea if his original idea was kids who post 
selfies on Facebook, but that wasn't the original showcase.

The LaTeX community back then already had found out that they needed styles.

> Why he based HTML on SGML instead of on LaTeX or the kind of markup
> used in Usenet postings is a good question.  HTML is worse than either
> for writing stuff manually, and that has led to stuff like WYSIWYG
> editors for HTML, which were pretty bad.  OTOH, TeX as a language is
> also badly designed, so I can understand that he did not want to go
> there.

IMHO, TeX already is one of my showcases, that you should not thing about 
adding a language late in the process.  Originally, TeX was supposed to do 
all the typesetting, and TeX, as programming language, just gradually 
evolved.

> Anyway, the Chuck Moore approach of throwing everything away upon
> encountering new requirements and designing the new thing from scratch
> was not taken for HTML.  Instead, the tried and proven methods of the
> software industry were used, and they result in bloat.  The web is 25
> years old, and has expanded in usage far beyond the original vision,
> so it's no wonder that browsers are bloated (the most popular ones are
> based on Mosaic from 1993, so even the software itself has seen
> "organic growth" for 22 years).  And Opera also has a 20-year history.

Hm, currently Chrome is the most popular browser, based on khtml, which was 
signifantly slimmer, and also easier to maintain, which was the reason why 
Apple picked it up to base Safari/WebKit on, and later Google joined. KHTML 
was started in 1999, 6 years after Mosaic, and the first usable version was 
released in 2000, as part of KDE 2.0.  At that time, all the stuff like CSS, 
DOM, and JavaScript had already been added, so Lars Knoll could design his 
code to the actual spec, which certainly is an advantage.

> And while users may moan about bloated software, they still buy it,
> and let the slim competition die from lack of revenue.

Ah, the browser industry is not a "buy" industry.  It's either free 
software, or deeply integrated in the operating system.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://bernd-paysan.de/
net2o ID: kQusJzA;7*?t=uy@X}1GWr!+0qqp_Cn176t4(dQ*

[toc] | [prev] | [next] | [standalone]

#36571 — Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?)

On Saturday, February 14, 2015 at 1:37:42 PM UTC-5, Bernd Paysan wrote:
> Anton Ertl wrote:
> 
> > Bernd Paysan <bernd.paysan.de> writes:
> >>You better ask Tim Berner-Lee why he invented HTML, and didn't just put
> >>LaTeX into his browser; the latter would have made a lot of sense, since
> >>all
> >>the people at CERN usually write their stuff in LaTeX anyways.  But it
> >>would have been more work upfront.
> > 
> > I think it would have been less work; the browser would have been just
> > wget (new, because it did not exist at the time. but simple to write),
> > latex and xdvi, with extensions for hyperlinks.  The problem would
> > have been that on the hardware of 1990 , it would have been too slow.
> > Our LaTeX benchmark takes 93s on a 486/66
> > <http://www.complang.tuwien.ac.at/franz/latex-bench>.  The best
> > hardware in 1990 was something like a 33MHz 486 or R3000, and
> > Berners-Lee had to target the majority of existing hardware, e.g.,
> > something like a 16MHz 386.  Ok, a web page would not take as much
> > time as our LaTeX benchmark, but still too much.
> 
> Berner-Lee only wrote for the NeXT cubes they had available.  Your benchmark 
> runs on a 140k byte LaTeX document, rendering 33 pages.  The typical web 
> page of that time was one page.
> 
> Furthermore, using LaTeX, you could have prerendered all pages, and just 
> transmitted the dvis.  xdvi really did perform very well, and was only 
> slightly larger than the original LaTeX (212kB for the 140kB of your 
> benchmark).  xdvi however was very slow on embedded images, since these 
> images were embedded as PostScript.  But that could have been solved easily, 
> just add another \special for gif and jpeg.
> 
> > Did they have to?  HTML 2.0 is fine for most of my uses, and 4.0
> > covers the rest (and provides lots of unneeded fluff).  I don't need
> > styles, and I am pretty sure they were not needed for what Berners-Lee
> > had in mind, either.  So he took the Forth approach: Created something
> > new tailored for his needs instead of catering for future expansion.
> 
> The original idea of Berner-Lee was to improve communication between 
> scientists at CERN.  The sort of people who already write a lot of LaTeX.  
> Maybe it would have been a bad idea if his original idea was kids who post 
> selfies on Facebook, but that wasn't the original showcase.
> 
> The LaTeX community back then already had found out that they needed styles.
> 
> > Why he based HTML on SGML instead of on LaTeX or the kind of markup
> > used in Usenet postings is a good question.  HTML is worse than either
> > for writing stuff manually, and that has led to stuff like WYSIWYG
> > editors for HTML, which were pretty bad.  OTOH, TeX as a language is
> > also badly designed, so I can understand that he did not want to go
> > there.
> 
> IMHO, TeX already is one of my showcases, that you should not thing about 
> adding a language late in the process.  Originally, TeX was supposed to do 
> all the typesetting, and TeX, as programming language, just gradually 
> evolved.
> 
> > Anyway, the Chuck Moore approach of throwing everything away upon
> > encountering new requirements and designing the new thing from scratch
> > was not taken for HTML.  Instead, the tried and proven methods of the
> > software industry were used, and they result in bloat.  The web is 25
> > years old, and has expanded in usage far beyond the original vision,
> > so it's no wonder that browsers are bloated (the most popular ones are
> > based on Mosaic from 1993, so even the software itself has seen
> > "organic growth" for 22 years).  And Opera also has a 20-year history.
> 
> Hm, currently Chrome is the most popular browser, based on khtml, which was 
> signifantly slimmer, and also easier to maintain, which was the reason why 
> Apple picked it up to base Safari/WebKit on, and later Google joined. KHTML 
> was started in 1999, 6 years after Mosaic, and the first usable version was 
> released in 2000, as part of KDE 2.0.  At that time, all the stuff like CSS, 
> DOM, and JavaScript had already been added, so Lars Knoll could design his 
> code to the actual spec, which certainly is an advantage.
> 
> > And while users may moan about bloated software, they still buy it,
> > and let the slim competition die from lack of revenue.
> 
> Ah, the browser industry is not a "buy" industry.  It's either free 
> software, or deeply integrated in the operating system.
> 
> -- 
> Bernd Paysan
> "If you want it done right, you have to do it yourself"
> http://bernd-paysan.de/
> net2o ID: kQusJzA;7*?t=uy@X}1GWr!+0qqp_Cn176t4(dQ*

I don't know anything about TeX, but I assume it is able to include pictures.

What about building a browser in Forth, using TeX, just for fun?

I know, it is a lot of work, but imagine:
We can use this special browser and start a community of Websites which are running on our Forth-TeX browser only. We can build up the whole stuff, and we don't need to follow any HTML rules - may be we need a wrapper of some sort.

What about that?
Running on Net2O or starting a Net2O?

Cheers,
Dirk.

[toc] | [prev] | [next] | [standalone]

#36572 — Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?)

visualforth@rocketmail.com wrote:
> I don't know anything about TeX, but I assume it is able to include
> pictures.

Pictures are added on TeX as an afterthought.  I used TeX as example of 
mature technology readily available in 1990, when Berner-Lee started.  It's 
not doing the things we expect a modern communication tool to be capable of.

Here are two screenshots how Berner Lee's browser looked like:

http://www.w3.org/History/1994/WWW/Journals/CACM/screensnap2_24c.gif
http://www.w3.org/MarkUp/tims_editor

Note that his bowser did have styles, just not as cascaded style sheets.

> What about building a browser in Forth, using TeX, just for fun?
> 
> I know, it is a lot of work, but imagine:
> We can use this special browser and start a community of Websites which
> are running on our Forth-TeX browser only. We can build up the whole
> stuff, and we don't need to follow any HTML rules - may be we need a
> wrapper of some sort.
> 
> What about that?
> Running on Net2O or starting a Net2O?

net2o's browser framework is already in my mind.  The graphics primitives I 
want to build on is OpenGL; the font technologiy is OpenType (so you can use 
freetype-gl to typeset), and you can integrate images and videos.  That's 
something we need today.

The main thing I want to highlight with LaTeX is that it has a typesetting 
programming language, open for the user.  This is what I want.  I don't need 
to carry the bloat of TeX with me, which is made for paper typesetting, and 
has the boundary between language and typesetting engine at the wrong place 
- IMHO, the typesetting language should implement the typesetting engine.

I.e. in net2o, you get fonts, OpenGL, images and videos, and the actual 
typesetting is up to you.  Of course, you are allowed to use "include".  But 
you have full control over the typesetting and rendering process.

That way I don't have much to think about the future.  I wouldn't start with 
LaTeX today, it also shows its age.  The typesetting language exists, but it 
is awful, as it was added as an afterthought.  Handling of UTF-8 has still 
rough edges, because the basic parser is designed for ASCII.  The boxes&glue 
model is sound, paragraph breaking and hyphenating algorithms are state of 
the art, so that's stuff to "steal".

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://bernd-paysan.de/
net2o ID: kQusJzA;7*?t=uy@X}1GWr!+0qqp_Cn176t4(dQ*

[toc] | [prev] | [next] | [standalone]

#36573 — Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?)

On Saturday, February 14, 2015 at 12:07:44 PM UTC-8, Bernd Paysan wrote:

> Pictures are added on TeX as an afterthought.  I used TeX as example of 
> mature technology readily available in 1990, when Berner-Lee started.  It's 
> not doing the things we expect a modern communication tool to be capable of.

What about the problem of image leaching then?

Jason

[toc] | [prev] | [next] | [standalone]

#36574 — Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?)

Jason Damisch wrote:

> On Saturday, February 14, 2015 at 12:07:44 PM UTC-8, Bernd Paysan wrote:
> 
>> Pictures are added on TeX as an afterthought.  I used TeX as example of
>> mature technology readily available in 1990, when Berner-Lee started. 
>> It's not doing the things we expect a modern communication tool to be
>> capable of.
> 
> What about the problem of image leaching then?

You mean the inline linking of images on other sites?  You see that as a 
problem?

net2o is a peer-to-peer network; so referring to any object (image or 
otherwise) is not costing you bandwidth - popular copies spread over the 
network.

If you want to preserve the integrity of your document, you could have 
embedded images, like in PDF.  The actual parsing of the document is 
something you can define yourself, so inline binaries which are decoded as 
images is entirely possible.

That's not something I want to do, because freely remixing and resharing 
content IMHO is an important function.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://bernd-paysan.de/
net2o ID: kQusJzA;7*?t=uy@X}1GWr!+0qqp_Cn176t4(dQ*

[toc] | [prev] | [next] | [standalone]

#36575 — Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?)

On Saturday, February 14, 2015 at 1:32:30 PM UTC-8, Bernd Paysan wrote:

> If you want to preserve the integrity of your document, you could have 
> embedded images, like in PDF.  The actual parsing of the document is 

That's just fine.

Jason

[toc] | [prev] | [next] | [standalone]

#36582 — Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?)

Bernd Paysan <bernd.paysan@gmx.de> writes:
>Anton Ertl wrote:
>
>> Bernd Paysan <bernd.paysan@gmx.de> writes:
>>>You better ask Tim Berner-Lee why he invented HTML, and didn't just put
>>>LaTeX into his browser; the latter would have made a lot of sense, since
>>>all
>>>the people at CERN usually write their stuff in LaTeX anyways.  But it
>>>would have been more work upfront.
>> 
>> I think it would have been less work; the browser would have been just
>> wget (new, because it did not exist at the time. but simple to write),
>> latex and xdvi, with extensions for hyperlinks.  The problem would
>> have been that on the hardware of 1990 , it would have been too slow.
>> Our LaTeX benchmark takes 93s on a 486/66
>> <http://www.complang.tuwien.ac.at/franz/latex-bench>.  The best
>> hardware in 1990 was something like a 33MHz 486 or R3000, and
>> Berners-Lee had to target the majority of existing hardware, e.g.,
>> something like a 16MHz 386.  Ok, a web page would not take as much
>> time as our LaTeX benchmark, but still too much.
>
>Berner-Lee only wrote for the NeXT cubes they had available.

I.e., a 25MHz 68030.

>Your benchmark 
>runs on a 140k byte LaTeX document, rendering 33 pages.  The typical web 
>page of that time was one page.

Yes, but it still would have taken too long on that hardware.

>Furthermore, using LaTeX, you could have prerendered all pages, and just 
>transmitted the dvis.  xdvi really did perform very well, and was only 
>slightly larger than the original LaTeX (212kB for the 140kB of your 
>benchmark).  xdvi however was very slow on embedded images, since these 
>images were embedded as PostScript.  But that could have been solved easily, 
>just add another \special for gif and jpeg.

Or transmit them as Postscript, which would have fit nicely with the
Display Postscript on the NeXT systems.  However, prerendering has the
disadvantage that it creates a fixed layout, not one adapted to the
window width and the font size preferences of the user.  There is a
reason why we don't use Postscript of PDF for web pages.

>> Did they have to?  HTML 2.0 is fine for most of my uses, and 4.0
>> covers the rest (and provides lots of unneeded fluff).  I don't need
>> styles, and I am pretty sure they were not needed for what Berners-Lee
>> had in mind, either.  So he took the Forth approach: Created something
>> new tailored for his needs instead of catering for future expansion.
>
>The original idea of Berner-Lee was to improve communication between 
>scientists at CERN.  The sort of people who already write a lot of LaTeX.  
>Maybe it would have been a bad idea if his original idea was kids who post 
>selfies on Facebook, but that wasn't the original showcase.
>
>The LaTeX community back then already had found out that they needed styles.

Scientific writers do not really need styles.  Hardly anybody who
writes something with LaTeX defines his own style.  We choose a style,
because we have to, but typically we either use the default "article"
style for articles, or a style the publisher prescribes.  That is
similar to the original HTML idea of the writer just providing the
content, and the reader deciding the style (which then consisted only
of the text width and font sizes).

The whole style nonsense came with the "web designers", who wanted a
specific layout and wanted to force it down the reader's throat.  That
idea is alien to both the original HTML and LaTeX.  Actually
Postscript and PDF are perfect formats for that, but somehow the web
designers and Adobe did not meet (at least not before Adobe bought
Macromedia).

>IMHO, TeX already is one of my showcases, that you should not thing about 
>adding a language late in the process.  Originally, TeX was supposed to do 
>all the typesetting, and TeX, as programming language, just gradually 
>evolved.

Yes, Knuth had expected people to take TeX and changing it for their
needs.  It turns out that even the style designers prefer to leave it
untouched, and build on top of it.

>> And while users may moan about bloated software, they still buy it,
>> and let the slim competition die from lack of revenue.
>
>Ah, the browser industry is not a "buy" industry.  It's either free 
>software, or deeply integrated in the operating system.

Modify "buy" as appropriate for the business model.  In this case, I
guess "use" is the right replacement.

- anton
-- 
M. Anton Ertl  http://www.complang.tuwien.ac.at/anton/home.html
comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
     New standard: http://www.forth200x.org/forth200x.html
   EuroForth 2014: http://www.euroforth.org/ef14/

[toc] | [prev] | [next] | [standalone]

#36585 — Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?)

Anton Ertl wrote:
>>> <http://www.complang.tuwien.ac.at/franz/latex-bench>.  The best
>>> hardware in 1990 was something like a 33MHz 486 or R3000, and
>>> Berners-Lee had to target the majority of existing hardware, e.g.,
>>> something like a 16MHz 386.  Ok, a web page would not take as much
>>> time as our LaTeX benchmark, but still too much.
>>
>>Berner-Lee only wrote for the NeXT cubes they had available.
> 
> I.e., a 25MHz 68030.

AFAIK the 1990 version already had a 25MHz 68040, and was "extremely fast" 
for that time.  For the color monitor, there was a i860 based card, which 
did the Display PostScript, and that was very impressive - the i860 did made 
a pretty good GPU.

>>Your benchmark
>>runs on a 140k byte LaTeX document, rendering 33 pages.  The typical web
>>page of that time was one page.
> 
> Yes, but it still would have taken too long on that hardware.

Some computer archeologists have dug out a fossiled NeXTCube, and demoe'd 
the first browser on the original machine:

https://www.youtube.com/watch?v=3c3Rt6QbHDw

It looks like rendering wasn't instantaneous, compared e.g. to opening the 
help windows, it takes a second or two to load and render a new page; though 
probably loading it was what took most of the time.

There are two reasons for the vision of Tim Berners-Lee not to use LaTeX: 
First, his browser had an interactive WYSIWYG edit mode.  That wouldn't work 
with LaTeX.  And second, there was a line mode browser for all those people 
who had just a terminal access.  Downgrading LaTeX to terminal mode is 
possible, but having a more limited format makes downgrading much easier.

> Or transmit them as Postscript, which would have fit nicely with the
> Display Postscript on the NeXT systems.  However, prerendering has the
> disadvantage that it creates a fixed layout, not one adapted to the
> window width and the font size preferences of the user.  There is a
> reason why we don't use Postscript of PDF for web pages.

From the look of the demo and the screenshots it seems to be that the first 
browser had a fixed width, and wasn't "responsive", as it is now called ;-).

>>The LaTeX community back then already had found out that they needed
>>styles.
> 
> Scientific writers do not really need styles.  Hardly anybody who
> writes something with LaTeX defines his own style.  We choose a style,
> because we have to, but typically we either use the default "article"
> style for articles, or a style the publisher prescribes.  That is
> similar to the original HTML idea of the writer just providing the
> content, and the reader deciding the style (which then consisted only
> of the text width and font sizes).

Ok, LaTeX styles come in two flavors: One flavor is the functional 
extension, and the other is design.  As content creator, you may need the 
functional extensions - if you don't need them, it would be a sign that 
LaTeX is not sufficiently modularized, and offers all the functionality you 
need by default.  Usually, you have 5 to 10 \usepackage{}s underneath the 
\documentclass just for a typical document.

The other is the "theme", the graphical layout.  That's not up to the 
content creator, but it is still necessary.  That's on the side of the 
publisher or reader.

> The whole style nonsense came with the "web designers", who wanted a
> specific layout and wanted to force it down the reader's throat.

That's a different problem: Apparently, those people were educated by MS 
Word.  As you say, they didn't even meet Adobe, which was the number 1 
address for commercial typesetting software.

> That
> idea is alien to both the original HTML and LaTeX.  Actually
> Postscript and PDF are perfect formats for that, but somehow the web
> designers and Adobe did not meet (at least not before Adobe bought
> Macromedia).

With the performance limitations of a 1990 high end machine, these pre-
formatted intermediate files indeed have something to offer.

>>IMHO, TeX already is one of my showcases, that you should not thing about
>>adding a language late in the process.  Originally, TeX was supposed to do
>>all the typesetting, and TeX, as programming language, just gradually
>>evolved.
> 
> Yes, Knuth had expected people to take TeX and changing it for their
> needs.

That doesn't work, because web is unmaintainable.  Nobody wants to change 
TeX.  Everybody wants to use the domain-specific language.

> It turns out that even the style designers prefer to leave it
> untouched, and build on top of it.

There's also the interoperability issue...  if you were to write a new LaTeX 
feature, and it requires your users to get your private fork of lualatex, do 
you think they would download it?  Probably not, especially not, if they 
want to use 10 such features which requires them to merge 10 forks of 
lualatex (with github, maybe they could even achieve that, but the 
likelyhood of a conflict is high).

But loading ten extensions written in that interactive interpretative 
language which TeX is into any of the currently four or five available TeX 
implementations, that works.

That means extensibility is not something that comes from releasing the 
source.  It requires structure.  Especially, it requires to put the 
interpreter or compiler of the extensible language right into your system.  
That makes it an extensible system.

If you look at Gforth, we have both Forth- and C-based stuff in there.  The 
build procedure of the C-based stuff is really, really awful, and complex.  
The build procedure of the Forth part is trivial, and that includes even the 
libcc.fs stuff, which has high internal complexity (we generate C code on 
the fly, compile it, create a shared library, load it and bind to the 
functions), but not for the build process itself.  It is trivial, because we 
have support of a powerful language.

Things only become hard if you have a limited special-purpose language which 
you can't extend.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://bernd-paysan.de/
net2o ID: kQusJzA;7*?t=uy@X}1GWr!+0qqp_Cn176t4(dQ*

[toc] | [prev] | [next] | [standalone]

#36601 — Re: HTML vs. LaTeX (was: Is 2015 the year of FORTH ?)

Bernd Paysan <bernd.paysan@gmx.de> writes:
>Anton Ertl wrote:
>>>> <http://www.complang.tuwien.ac.at/franz/latex-bench>.  The best
>>>> hardware in 1990 was something like a 33MHz 486 or R3000, and
>>>> Berners-Lee had to target the majority of existing hardware, e.g.,
>>>> something like a 16MHz 386.  Ok, a web page would not take as much
>>>> time as our LaTeX benchmark, but still too much.
>>>
>>>Berner-Lee only wrote for the NeXT cubes they had available.
>> 
>> I.e., a 25MHz 68030.
>
>AFAIK the 1990 version already had a 25MHz 68040

Yes, but <http://en.wikipedia.org/wiki/NeXT_Computer> says that
Berners-Lee used the 68030 version.

>There are two reasons for the vision of Tim Berners-Lee not to use LaTeX: 
>First, his browser had an interactive WYSIWYG edit mode.  That wouldn't work 
>with LaTeX.  And second, there was a line mode browser for all those people 
>who had just a terminal access.  Downgrading LaTeX to terminal mode is 
>possible, but having a more limited format makes downgrading much easier.

TeXinfo would have been another candidate, very much in line with that
requirement.

>> Scientific writers do not really need styles.  Hardly anybody who
>> writes something with LaTeX defines his own style.  We choose a style,
>> because we have to, but typically we either use the default "article"
>> style for articles, or a style the publisher prescribes.  That is
>> similar to the original HTML idea of the writer just providing the
>> content, and the reader deciding the style (which then consisted only
>> of the text width and font sizes).
>
>Ok, LaTeX styles come in two flavors: One flavor is the functional 
>extension, and the other is design.  As content creator, you may need the 
>functional extensions - if you don't need them, it would be a sign that 
>LaTeX is not sufficiently modularized, and offers all the functionality you 
>need by default.

One has nothing to do with the other.  Modularization is about the
internal organization of the program, the default functionality is
about user convenience.  It is often useful to expose the modularity
to the users to support extensibility, but that does not mean that the
user has to see the internal modules.  E.g., Gforth does not become
more modular if we don't have float.fs or history.fs loaded by
default, only less convenient.

>Usually, you have 5 to 10 \usepackage{}s underneath the 
>\documentclass just for a typical document.

And the \usepackages are typically about allowing the author to
express content, not about layout.  The \documentclass is about
layout, but is obligatory.

>> The whole style nonsense came with the "web designers", who wanted a
>> specific layout and wanted to force it down the reader's throat.
>
>That's a different problem: Apparently, those people were educated by MS 
>Word.

Or the MacIntosh and its WYSIWYG paradigm.

>> Yes, Knuth had expected people to take TeX and changing it for their
>> needs.
>
>That doesn't work, because web is unmaintainable.  Nobody wants to change 
>TeX.  Everybody wants to use the domain-specific language.

Yes, his expectations were unrealistic, but they explain why he did
not design the programming language aspect of TeX more diligently.

- anton
-- 
M. Anton Ertl  http://www.complang.tuwien.ac.at/anton/home.html
comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
     New standard: http://www.forth200x.org/forth200x.html
   EuroForth 2014: http://www.euroforth.org/ef14/

[toc] | [prev] | [next] | [standalone]

Page 1 of 3  [1] 2 3  Next page →

Back to top | Article view | comp.lang.forth

csiph-web