I just have a few things to say to purge my opinions having finished working with SML (for good).

SML isn’t a bad language. It’s (mainly) functional and if you like functional programming or think in a certain way you will probably like it. It’s very elegant and neat, things happen in a few lines and you can do some very clever stuff.

The best things I liked about SML were:
-the map function: you write a function that applies to one element of a list and the map function will give you a function that applies that to the whole list
-curried functions and partial instantiation: also quite handy. Curried functions allow you to partially instantiate things (i.e. you can pass parameters one at a time (which is a great deal more useful than you would think).

I think it’s good points are also it’s bad points. I don’t care about elegant code and hence it infuriated me when my project supervisor went all starry-eyed about how wonderful SML is. The fact you can do stuff in a few lines is very cool; it’s also very confusing. Setting something out in a longer-winded way can often lead to code that a larger number of people can read and understand.

One of my other problem with SML was that I had to learn it before I could use it. This inevitably meant that I didn’t use the fancy stuff that SML can do but something like C++ (or dare I say it Python (think higher-order functions other than map)) can’t.

Now having finished talking (well…mainly complaining) about SML as a language – I have a few other problems.

1) The documentation is crap. It’s fairly non-existent. Most of the stuff on the web is also pre-1997 when there was a big change in the language and hence it’s useless. Nobody tells you this though! If anyone reading this has to learn to program in SML – buy a book

2) You can’t do anything with it – No-one writes APIs (interfaces to programs) for SML. Also there are two systems available to make GUIs – both are appalling, particularly the ones the SML website wants you to use. They both appear to be ‘proof of concept’ rather than anything anyone is expected to use. Anything that is written for SML (including the Gui systems) has the documentation issue too (see 1)

3) It’s only used by academics – perhaps this is the root of all the other problems. In essence, nobody writes about it, nobody uses it and even if somebody did use it they couldn’t do anything with it.

Basically my problem with my project was that I see no point in using the language I was being forced to use. I didn’t use the fancy stuff and that is the only reason to even think of using SML

Sorry for the the long rant. It’s my last on this subject I promise. Happy programming, live long, prosper and don’t use functional languages.

Actually, while I think about it. The documentation is crap. If anyone trying to write in ML actually reads this. Ignore most online tutorials as they were all written before 1996 and so don’t actually bear any resemblance to the language nowadays. Nobody tells you this. That took me ages to find out

I promise that in a few days I’ll stop moaning about SML and get on with my life. I may however write up some thoughts and tips just incase some other poor sap being forced to use it finds it and makes use of it.

I have a few more things to code up and I’ve got a report to write. It’ll all be ok as long as I work like a maniac. Two things concern me however. Firstly I’m a little concerned whether the project is complicated enough. It hardly seems like a work of genius at the moment but maybe I’m being pessimistic. Secondly, and the less abstract of the two is I have visions of getting into college printing it out and it looking completely different to the way it’s meant to. It all comes of not owning a printer I guess.

Oh well, to distract me today I have the Physics Ball to help with.

As long as I get a 2:1 I don’t care (to a certain extent)

> hello!
> i have think how to do something like this:
> for example i have
> [1,3,2,1,2,4,2,1,3,4]
> and i want to get
> [1,3,2,4]
>
> The idea is to delete repeted elements.
> can someone help me?

The response was

“Recursive thinking”:

I know how I’d write a function to do that and I don’t think that response was in the least helpful. Infact I’m now slightly more confused. It really does make me wonder sometimes.

I’d also like to see some examples for the graphical system I’m trying to use. Lots and lots of people appear to be talking to it but whenever you google for it you just get a load of pages linking to the project homepage. I’m beginning to doubt whether anyone ever uses sml_tk. I think all they do is talk about using it. All I want to know is how I call an external function. Another thing I’ve noticed is lots of people calling about 2 A4 pages of function listings ‘Documentation’. grr

I’ll stop being angry now. On the plus side I’m getting into this functional programming lark now.

So much to do and so little time. Which is obviously why I’m wasting my time writing this…..

Oh well, at the moment I can’t complain. It all seems to be going ok. (famous last words)

Ahahahahahahahah!!!!

I’ve actually made some progress. My program will now read in an FSM (really must re-write the read in function so that it isn’t a fixed file name), do one or two really minor things to it (guess what my next task is) and then write it out in a format the a graph drawing program can read, run the program and then run a viewer.

I’m chuffed. Still loads of work to go but I’m getting there. It does something now.