Open Turing (How to make turing)

[wtd]

Tony wrote:

Freakman wrote:

we should try and keep it as close to the official versions speed as possible. (That is, if this IDE thing gets off the ground)

But that means intentionally crippling the language (with what? artificially implanted clock-cycle caps?)

I just see no appeal to an open source Turing clone beyond where Turing is currently used. As such, compatibility in all aspects with that existing environment is the top priority.

[Cervantes]

Tony wrote:

We'd just have to use another dead Mathematician's name in place. (You know, the whole IceWeasel deal)

How about Alonzo Church? He's closely linked to Turing (see Church-Turing thesis) and fathered lambda calculus.

I'd say we shoot for a CLI first. Once that's been done, the road to an IDE is open.

[Tony]

Quote:

[Alonzo Church] and Turing then showed that the lambda calculus and the Turing machine used in Turing's halting problem were equivalent in capabilities, and subsequently demonstrated a variety of alternative "mechanical processes for computation."

Cervantes, I am awed with your brilliance

[rdrake]

[[Gandalf]]

Wow, nice rdrake. You forgot syntax highlighting though.

wtd wrote:

I just see no appeal to an open source Turing clone beyond where Turing is currently used. As such, compatibility in all aspects with that existing environment is the top priority.

Oh dear... For once I must strongly disagree.

I mean, yes, a huge part of the reason for making this is to help kids out with school projects and homework, but that's hardly reason to do something as insane as purposly crippling something. Leave that to the commercial programmers...

First off, it would be possible to leave it at full speed by default, and then have an option to either put in a delay or else waste clock cycles on 'nothing'. Same thing goes for extending the actual language capabilities or library, have a backwards compatible mode, and then an extended mode.

Second off, you'd never get the exact, or probably even close to the real speed of each function of Turing. It's not worth the effort it would take. Imagine all the things that affect the execution time of some code, who feels like wasting time measuring how long each function in Turing takes to run in a whole bunch of test cases? Not I...

Third off, letting problems like "do we make it slow when it doesn't have to be?" cripple the beginnings of this project is the first step to it dieing out, like everything else.

In the end, you can only go so far in "compatibility in all aspects."

[zylum]

why would you want to make it slower than necessary? i would have to disagree with that point. but we should make the language identical atleast for the first version. then if we make newer versions we can depricate some stuff and add in our own functionality.

[md]

Compatible in the sense that syntax and stardard library functions are all the same. Beyond that anything new is an extension, and the only good ones I can think of are linking turing function headres to pre-written code in another language. That allows you to import all the additional features you want if they use non-turing-standard APIs (like say, OpenGL).

[[Gandalf]]

Good idea md, though extending actual language features might be a bit interesting too.

As for what I said, I assumed that wtd was also talking about crippling the speed, since he replied to:

Quote:

But that means intentionally crippling the language (with what? artificially implanted clock-cycle caps?)

In which case I fully agree with zylum.

[wtd]

Perhaps a few questions should be answered up front.

What is the target audience for "Open Turing"?

What problems does that audience have?

How will it solve those problems?

[Andy]

i brought up an interesting idea on another topic. what about hooking our turing compiler up with CGI and allow it access via the internet, and make an php turing editor?

[Cervantes]

I don't see much of an advantage to that. Instead of asking students to download and install and run the program, they have to go online and load this webpage. Also, it would be a lot slower.

[Andy]

there are several advantages.

1, editor would be extremely easy to build, since we already have turing syntax highlighting implemented into compsci.ca. 2, it would be able to determine what system the user is on and automatically build the required binary for the user, without having them worry about which version of openTuring to install. and finally, it would give us a chance to track bugs within the software, if the program crashes, we'd know it and have memory dumps of what went wrong.

[Cervantes]

1) If we're going for a CLI, this doesn't even apply. And I would imagine the syntax highlighting for compsci.ca could be ported to whatever IDE we build, even if that just means copying and pasting a huge array of keywords or whatever it may be.

2) I wouldn't forsee a problem of students downloading the linux version of openTuring and trying to install it on their windows machine, or (especially!) vice versa.

3) Good call here.

[md]

forget the GUI aspects for now, though I too think a web-based gui is mostly pointless.

In answer to wtd's questions:

Quote:

What is the target audience for "Open Turing"?

I think there are actually two audiences. The first is people who wish to write turing code at home, and don't want to resort to pirated copies of Turing. This group may not run windows, and they may or may not be studying comptuer science in some schooling environment. This would most likely be the larger group of users.

The second, smaller, group are those (like me) who see the occasional turing program which looks interesting, but have no way of running it to check it out. Again this group may not be running windows. However unlike the first this group is unlikely to do very much in the way of actual programming, excepting the occasional debugging.

Quote:

What problems does that audience have?

The problems faced are again two-fold. In the absence of a free implementation of a turing compiler anyone wishing to run programs must find and download a copy of Turing. This is in violation of copyright law; and is not a legally or ethically valid choice. While many people have done this, I would think that were a sufficiently easy to use compiler made available for free they would use it instead of resorting to copyright infringment, and potentially theft (depending on the interpretation of copyright law).

In addition to the lack of a free turing compiler, anyone wishing to use Turing, legitimately or not, is forced to run Windows or MacOS in some fashion. This again limits a potential users options as if they do not run Windows/MacOS - or worse do not own either - they are forced to install (again perhaps an illegally acquired copy) of one or the other.

As the majority of Turing users are students, placing the added burden of running a non-free operating system, on potentially non-free hardware, will almost certainly place them in a position where they must violate copyright law in order to run their own turing programs. This is not acceptable.

Quote:

How will it solve those problems?

A free turing compiler solves these problems by removing the limitations of the commercial Turing compiler. Specifically it is free, so students can easily afford it, and if the source code is released under an open source license (and I firmly believe that should be the case) then it could easily be ported to other systems such as linux or *BSD. That further reduces or removes the burden of running a non-free operating system such as Windows or MacOS by not requiring the use of such operating systems.


The next question is "How do we get there?" I think that is where things will get messy.

[Windsurfer]

Can we call it openFlugge-Lotz? (just because i find the name funny)

Or, perhaps call it openEinstein? After all, he did once say "I don't believe in mathematics." And it would be entirely ironic.