Message Boards Message Boards

9
|
28455 Views
|
11 Replies
|
19 Total Likes
View groups...
Share
Share this post:

OOP environment for Mathematica without additional packages

Posted 9 years ago

I'd like to introduce the OOP environment for Mathematica, with no-additional packages.

Mathematica has many style of paradigm, however, the programming style of OOP is an old but not sufficient backed-up environment. Let's consider the function group UpSet. UpSet has been seemed the basic idea relating OOP and the function group being equipped from the first version, but you can see the description of the function group is so small and short in the manual.

Now, the UpSet function becomes the most important function because the function group is the key parts to establish the OOP environment. With very small size pattern of CLASS constructed with SetDelayed UpSetDelayed, and Module leads to the OOP paradigm. This paradigm includes Encapsulation, inheritance, and Polymorphism in very straight forward style. OOP on the Mathematica suits for Graphics, Events, and the Dynamic directly and give us new kind of power for Mathematica programming.

The method is very easy as follows. Basically, you define the CLASS which is "SetDelayed definition of Module-d UpSetDelyaed functions", you construct INSTANCES distinguished by the name having each local values, and each INSTANCE functions when you send name conbined MESSAGES defined in the CLASS. In the following code, class name is "pclass", two method named "set" and "get", object list name is "object". In my definition, OBJECT is the group name of INSTANCE created from CLASS.

(*prepare class*)
pclass[nam_] := Module[{local = 0},
  set[nam[x_]] ^:= local = x;
  get[nam] ^:= local^2;]
(*prepare object list*)
object = Table[Unique[], {10}];
(*create instances*)
Table[pclass[object[[i]]], {i, 10}];
(*send message to set local values of instances*)
Table[set[object[[i]][i]], {i, 10}];
(*send message to get results of instances*)
Table[get[object[[i]]], {i, 10}]

{1, 4, 9, 16, 25, 36, 49, 64, 81, 100}

I hope you enjoy the new paradigm.

In Google slide, you can see more detailed OOP discussion.

Attachments:
11 Replies

It would be great I think if you would be able to explain your OOP idea in-post here or attache a notebook.

POSTED BY: Sam Carrettie

Yes, I attached a notebook.

Posted 8 years ago

Hi,

Apologies, yet I do not see how your approach promotes re-use, abstraction, encapsulation, or Typing of any sort. Despite your mention of design patterns, I do not see how it can, for example, be used to implement common patterns such as observer, visitor, or, most ubiquitously, singleton. Most importantly, I unclear how it helps me, in structuring code this way, reduce complexity or improve understanding.

I did download your notebook. Unfortunately, this was not enough to actually understand what it is that you are proposing and what advantages there are in contrast to other approaches. I think much might be understandable were there a more in-depth write-up similar to that others have posted when putting forth their proposals for approaching OO in Mathematica.

Regards, Dave

POSTED BY: David Fuller

You can add something in the general topic on SE or in topics linked there:

Which Object-oriented paradigm approach to use in Mathematica?

enter image description here

POSTED BY: Kuba Podkalicki

This mind map is asking to reconsider on the Mathematica OOP, "how to" develop or/to "why to" develop it. Though, if we can achieve OOP with very small cost, the importance of "why" becomes lower one. My idea for the OOP requires very small cost, it means a few lines code is additional, so "why" will have small meaning in my OOP case. On the contrary, we can obtain large advantage from the OOP.

The power of Mathematica comes from many characteristic points, and I think the various programming paradigm may be the largest one, so OOP will become the new large power source.

Why name the function Upset? Is it a cute metaphor for the pain of programming OOP in general as well as describing how parameters are set up for inheritance? If so it worked. I got a good chuckle.

POSTED BY: Jules Manson

Je ne sais pas. Please ask to Stephen Wolfram, why UpSet?

Posted 8 years ago

One more Q: Why would I 'construct' a superclass specifically that I might 'subclass' it? That does not appear to make sense. I am assuming that your example is just unclear in some respect.

POSTED BY: David Fuller
Posted 8 years ago

So you are asking why is a superclass necessary in general? Briefly, it is related to the design pattern and to make the design looks clear and logical.

POSTED BY: Shujie CHEN
Posted 8 years ago

No, I'm not asking in general. I well understand the purpose which is why I asked the questions I did.

I'm asking why, apparently something must be declared to be a superclass using a super function? Is that not the point of stating that the child generalizes the parent?

This is why I'm asking for some more exposition regarding this approach. If it, somehow, provides the benefits of OO simply by convention then it is indeed well worth understanding how exactly it does this and what positive and negative effects may accrue.

POSTED BY: David Fuller

OOP, Object Oriented Paradigm may not be familiar as a tool for Mathematica programming, so, I'd like to introduce some examples using OOP, considering this six months discussion after the publish of my idea.

Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Reply Preview
Attachments
Remove
or Discard

Group Abstract Group Abstract