[Axiom-developer] Re: Stylesheet

[Arthur Ralfs]

root wrote:
>> One of my longer term interests is more network awareness and parallelism
>> with Axiom.  For instance I imagine wanting to run axiom on a powerful
>> server and then connect to it from some other computer.  It might just
>> be that I'm running Axiom on my own server and connecting to it with my 
>> laptop.  In >that case though  don't we need to go through the server?
>>     
>
> When I have a quad or 16-way processor on my laptop I'm not sure what
> benefit a "powerful server" adds. I'm not even sure what "powerful
> server" might mean. Perhaps you could give more details about your 
> longer term view?
>   
That's true that laptops can be very powerful now-a-days however
desktops still give much more computing power for the buck.
Certainly running locally is fine in lots of cases but I would still like to
have the option of running over a network.

In my experience no matter how powerful the computer I can quite easily
bring it to a near standstill with a computationally intensive project.
For instance in the classical N-body problem it's just a question of making
N big enough.

> I do see parallelism being important to Axiom as I'd like to be able
> to handle both parallelism-in-the-small, e.g. matrix multiply and
> parallelism-in-the-large, multiple problem solutions or multiple
> non-algorithmic procedures, like magnus, which are not guaranteed
> to terminate. But I don't see the need for remote servers anymore.
>
>
>
> Another advantage of a local axiom is that I can see a person working
> on Axiom who does not know the syntax of a function (say, a laplace
> function call). In the middle of their session they could do )help
> laplace which would put them into a screen to construct the function
> call (see differentiate and matrix). They would construct their
> function call and it would be added to their workspace as it is under
> the current setup. Once they quit the )help they have already defined
> the function call. 
>
> Alternatively a user could make a new frame (see the )frame command)
> and do all of their )help browsing in a separate frame namespace so
> it does not impact their current work.
>
> The frame mechanism allows Axiom to support multiple namespaces of
> work and could probably be applied in a server environment.
>
>
>
> That said, I don't see any disconnect between the current test
> setup and a remote server. Instead of doing 
>   file:///...../rootpage.html
> you could certainly do
>   http://....../rootpage.html
> In fact, that would probably work now although I haven't tried it.
>   

As long as either option works I'm happy.
>
>
> If we want to widen this discussion perhaps we should journal it
> on the axiom-developer mailing list.
>
> Tim
>
>