MathGroup Archive 2008

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Re: Range of Use of Mathematica

  • To: mathgroup at smc.vnet.net
  • Subject: [mg89058] Re: [mg88958] Re: Range of Use of Mathematica
  • From: Andrzej Kozlowski <akoz at mimuw.edu.pl>
  • Date: Sat, 24 May 2008 03:58:17 -0400 (EDT)
  • References: <g0m8tt$14$1@smc.vnet.net> <g0rkfr$dtv$1@smc.vnet.net> <200805200627.CAA23235@smc.vnet.net> <g11qvh$a6r$1@smc.vnet.net> <200805220637.CAA22441@smc.vnet.net>

On 22 May 2008, at 15:37, David Bailey wrote:

> Andrzej Kozlowski wrote:
>> On 20 May 2008, at 15:27, David Bailey wrote:
>>
>>> One sign of bloat, is that there are now two major interfaces -
>>> notebook
>>> and Workbench (not to mention Math.exe) and such a profusion of
>>> commands
>>> and options that even those of us that use the product regularly
>>> cannot
>>> hope to be familiar with more than a smallish subset. There are also
>>> two
>>> help systems! This has, unfortunately, impacted on the quality of
>>> documentation. My advice to a beginner would be to read the V5.0 or
>>> even
>>> V4.0 documentation (plus a few tips about the changes in Graphics)  
>>> to
>>> get some idea of the way the system works.
>>
>>
>> I think one should not loose sight of in my opinion, a  very  
>> essential
>> fact, which is that the Mathematica Kernel, has remained largely
>> unchanged, except (of course) for continual enhancement by addition  
>> of
>> new functions, which only make it more efficient and powerful (and
>> make programing considerably easier). What this means is, that people
>> who see themselves essentially as Kernel users, are not in any
>> fundamental away affected by any of the changes you are mentioning
>> above. I know, because I consider myself one of them. I have never
>> tried using Workbench (even though I have downloaded it), and I have
>> not even made a serious effort to learn how MakeBoxes etc, works. I
>> still write all research in TeX. I know that it is possible to use
>> FrontEndToken etc, and if I even wanted to do so I know where to find
>> the documentation - but I have never found any incentive to do this.
>> However, I have become interested in the new Dynamic functionality,
>> partly because I think it is a great teaching tool and partly because
>> of the  Demonstrations project, but this is really the first new
>> addition to the Front End that has interested me. I am sure that I am
>> not alone. In fact, Wolfram has a different set of people working on
>> the Kernel and on the Front End, and I am sure that there are people
>> in the Kernel group whose knowledge of the workings of the Front End
>> isn't much greater than mine.
>>
>> I mention this because even though I have not, until recently, been
>> very interested in the Front End, I have never been in anyway
>> inconvenience by the developments that have taken place in that area.
>> True, Mathematica has grown larger, but not more than have my hard
>> disks or RAM. I therefore, see none of the problems you mention or
>> that seem to worry AES. I am a mathematician and intend to remain  
>> one.
>> I use Mathematica in the way that it was originally advertised, as a
>> "system for doing mathematics by computer". The most far reaching
>> change in the program that was ever made happened, in fact, in  
>> version
>> 2, when the Kernel was separated from the Front End. Sometime little
>> later WRI changed its advertising slogan to something like  "the
>> world's only fully integrated technical computing system". That, of
>> course, shows clearly the strategic direction that WRI chose and it
>> has consistently followed since. But for me it has always remained "a
>> system for doing mathematics by computer". In spite of that, I see no
>> reason to complain, because I have never found the slightest conflict
>> between these two "roles" of Mathematica. I have never found , for
>> example, that the development of Mathematica's typesetting
>> capabilities has in any way adversely affected any of Mathematica's
>> numerical or algebraic ones, which are the ones that really matter to
>> me.
>> So I really still fail to see what this whole discussion is supposed
>> to be about. I particularly, I can't understand why someone who keeps
>> saying that Mathematica does not need publishing or presentation
>> capabilities etc. and should only be used for computations would at
>> the same time complain about the supposed lack of documentations of
>> functionality which, according to him,  should not be there in the
>> first place.
>>
>> Andrzej Kozlowski
>>
>>
>>
> Obviously, WRI are very reluctant to remove functionality, so our code
> goes on working from version to version. That is great, but we surely
> also want newcomers to find Mathematica easy to use.
>
> Unless a new user decides to go on a course - with all the extra  
> expense
> that involves - they have to try to get an overview of the system and
> decide which features are important to learn about. I suspect that  
> this
> is not as easy as it may seem to those of us that have used  
> Mathematica
> for years, and know how it all fits together.
>
> The new documentation system encourages a browsing style of learning
> that may miss out vital ideas. For example, a user could be forgiven  
> for
> deciding not to invest time learning about the relationship between  
> the
> FE and the kernel. As a result he may fail to understand how code in
> different notebooks interoperates.
>
> API's can get too big - try programming in the WIN32 API!
>
> David Bailey
> http://www.dbaileyconsultancy.co.uk
>

I think you somewhat misunderstood my point (and I think AES's point  
too). To put it in a nutshell: if you wish to use Mathematica only as  
a computational engine and use TeX, Illustrator and whatever other  
programs you wish to use for non-computational tasks, then you can do  
this just as well with Mathematica 6 as with 5 or earlier, and in fact  
much better, since the computational engine is now much more powerful.  
It has become much more powerful without any fundamental changes to  
the basic structure of the language (which is not the case with some  
other rival programs). So, if you do not care for Mathematica's  
typesetting, graphic and presentational capabilities you can simply  
ignore them - they in no way affect the computational ones. Indeed,  
this has to be so, otherwise the Kernel would not be able to function  
without the F.E. and it obviously does.
Your concerns about the FE becoming over complex for newcomers  
(whether justified or not) affect only those newcomers who need  
capabilities other than pulling down menus, copying and pasting and  
using Shift-Enter. If I am not mistaken, AES thinks that these are the  
only FE capabilities Mathematica needs. My point is that for users who  
adopt his approach (which I have nothing against in itself,  
particularly that it isn't all that different from mine) there has  
been no significant in difficulty and complexity. Even the complaint  
about the lack of printed documentation does not carry much weight, if  
all that you care about is the Kernel, since available documentation  
for earlier versions is more than enough to turn one into a (Kernel)  
power-user.

Finally, I would like to add one comment on the issue of WRI's  
"strategy" that AES has raised. Actually, this strategy has been very  
clear many, many years, in fact even since WRI decided to re-position  
Mathematica from a "system for doing mathematics by computer" (in  
other words a "CAS") to "fully integrated technical computing system".  
Note the words "fully integrated". To me it has been clear for more  
than a decade that this was meant to imply things like typesetting, a  
fully programmable graphic engine (independent of PostScript), and  
other things. In fact, any one who attended any of the numerous  
Mathematica demos, particularly those conducted by Theo Gray, could  
have had no doubt about this. Has this been a success? Well, let me  
point out that quite a large number of commercial general-purpose CAS  
have appeared since Mathematica's debut (and some even earlier). With  
one notable exception none of these has been a commercial success and  
most are available now as freeware. I would say that is a pretty clear  
vindication of Wolfram's strategy.
By all means have a discussion of the "product strategy" of  
Mathematica if you enjoy this sort of thing. Butif if anyone hops to  
stop or reverse the ongoing process of "integration" (as in "fully  
integrated") than all I have to say is that it seems to me to be  
rather late for that; by over a decade, actually.

Andrzej Kozlowski


  • Prev by Date: Re: "Reduce" wierdness (or too slow?)
  • Next by Date: Re: "Reduce" wierdness (or too slow?)
  • Previous by thread: Re: Re: Range of Use of Mathematica
  • Next by thread: Re: Range of Use of Mathematica