On 04.04.2012 14:05, Ashwini Oruganti wrote: > On Sat, Mar 31, 2012 at 11:37 PM, Alexey U. Gudchenko <pr...@goodok.ru>wrote: > >> >> On 31.03.2012 14:02, Ashwini Oruganti wrote: >>> Hi, >>> >>> My proposal can be found on the SymPy wiki at : >>> >>> >> https://github.com/sympy/sympy/wiki/GSoC-2012-Application-Ashwini-Oruganti:-Step-by-Step-Expression-Manipulation >>> >>> Can you please review the same and provide feedback? >>> >>> >>> Thanks and Regards, >>> >>> Ashwini Oruganti >>> >> >> 1. As I understand, there are two main parts of your application: >> >> a) informational: to inform what SymPy do by steps, >> b) control: manage with those steps. > > (exclusive of interface design) >> >> I think that those tasks must be clearly separated by the aims >> (nevertheless they are have the same base, an one of them use the other). >> > > I realize that you and I are saying almost the same things, but under > different subheadings. I tried to divide my design into the UI part and the > main algorithm designing part to generate steps, as I think it's easier to > explain the design using this. > > > >> >> 2. We must think how it can be embedded into SymPy, and easy maintain >> synchronization with changed code. >> > > As I'm new to SymPy, I plan on doing this over along the way as I code. In > my proposed schedule, I've allotted some time for such design discussions. > I'm afraid my current knowledge about SymPy isn't enough to exactly answer > this question. I would be happy to include any inputs you might have > regarding the same. > > >> >> As I understand you are going to create the set of rules (Data >> Structure) as the central part. >> >> If the algorythm of the solve function (integrate|simplify|...) is >> updated we must to make adjustments to the rules too. >> >> The typical function: >> >> solve(expr): >> try to solve by method1 >> try simplify >> try sub_method_1 >> if ok: >> return >> if failed: >> try to solve by method2 >> >> What is the DataSet for this example? >> > > IMO the data-set will be different for different operations (e.g. For > differentiation, it will be product-rule, first degree derivative, quotient > rule, etc). > > > >> >> I think, Most likely it will follow much of the code. >> >> May be to think how will the "rules" be created automatically? E.g. with >> the help of statements embeded in the code like e.g. >> 'steps.push('info')" "cancel_transaction" functions? >> >> For the task (a) it can help to maintain the actuality of steps with >> respect to the present code. And even if the one of steps description >> will be missed the code (of step-by-step manipulation module) doesn't >> crash. >> > > Yes, for this purpose I plan to include smart defaults in my algorithm. > Exactly "how", you might ask, but I'm working on it. I might be a little > slow here, but I'm trying to figure ways out. That's the reason why my > timeline has a lot of time for designing and discussions with mentor. > >
I think, the starting point is to create imaginary variants of the code examples of such procedures, like that is written above but with real sympy expressions in input and output for every step. This will be "original" code. Then invent the ways how to join it with step-by-step: what is needed for developer of this code (whom don't know much about "step-by-step"), and simulate how "step-by-step" works with this examples. -- Alexey Gudchenko > >> For the task(b) this way must be changed somehow, but the task (b) can >> use the information collected in task (a). >> > > > Indeed, that's what I've tried to put forth in my proposal. I guess I'll > add a flowchart of the design so as to make it more clear. > > >> -- >> Alexey Gudchenko >> >> -- >> You received this message because you are subscribed to the Google Groups >> "sympy" group. >> To post to this group, send email to sympy@googlegroups.com. >> To unsubscribe from this group, send email to >> sympy+unsubscr...@googlegroups.com. >> For more options, visit this group at >> http://groups.google.com/group/sympy?hl=en. >> > -- You received this message because you are subscribed to the Google Groups "sympy" group. To post to this group, send email to sympy@googlegroups.com. To unsubscribe from this group, send email to sympy+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/sympy?hl=en.
