Re: [Scilab-Dev] Q/A about features
Hi Amanda, Le 23/02/2017 à 12:59, Amanda Osvaldo a écrit : .../... (Save the environment) Yes, in the client I can do it. But how to do in the CLI in text mode? It's to automate many operations and to expand the work in text mode. You shall use: save mySessionFile load mySessionFile * Interaction with users Yes, including I have reported issues in bugzilla. But I don't know if a user takes more than 2 minutes to be able to report something, I believe he can desist. The application has no function to guess or read users minds and post a report accordingly. It has not yet any embedded A.I. to do so. * % keyword as comment: The need it's simple. Simplicity is not necessarily relevance nor justification. Not always a software can do everything. O wise words :) So to improve the job sometimes it needs to switch to a software to other in differents steps of the job. Think about the differences in modules, hardware consumption, and cost. In common they always will have the same scripts. Obviously not. Other language, other scripts. So better compatibility make this exchange smoother. For example: I can convert images to HDF5 with Octave. Run a personalized module in SciLAB Make the heavy job in a cluster with MatLAB. And Finish in SciLAB with the personalized module. But just, Matlab and Scilab are not compatible. Adding % comments would not make then compatible. It is not a matter of more or less compatibility. And no always what macro will run in each software it's clear. So it's necessary that the 3 share the same scripts. So, you wish to build a kind of computing esperanto. Could you please give us a pointer to your wish certainly posted on Octave mailing list or bug tracker to support "//" as comment as well? Look at this Octave session: >> // a comment parse error: syntax error >>> // a comment ^ What a shame! And a potential loose of manpower! Don't you think? And try with "#" that is another symbol used for comments in other widespread languages. The same! So bad is the world. * Colors and arrays with names instead of numbers Is about to reduce the number of steps and gain productivity. Including for more complex selections. Using the %nan i need: 1. to write the selection operation 2. make a array copy 3. put %nan in the selection 4. display the array 5. restore the original array It's 5 steps. :-O If the selection it's always the same so why do not reduce the visualization to a only single step ? Repeat this 100 times on the same day it's a nightmare. And more, the idea to put colors and words can be used to introduce sense in the numbers. A matrix may have not only one word, but many. For example: Normal, Caution, Danger, Error. The colors can have the same meaning. I don't know If I'm being clear It was not very clear the first time i read (and reread) it, but now it's ok (likely). In: --> consoleSetProfile(“my-color-profile”); --> disp(data) 1. 2. 20. 1. 2. 17. 7. 6. 18. the most important is consoleSetProfile(“my-color-profile”); where the prerequisite is to have set your specifications (i guess). Excepted for the colors, you can do this in Scilab by defining your own Disp() that will preprocess your input data before calling disp() or mprintf() to actually display them. About colors in the console, your skills and time are certainly welcome to process this report: http://bugzilla.scilab.org/10142 ** Pool of formulas* It's a dilemma. It I have a good engineering team but they can not or do not know how to make a good SciLAB implementation, I will need another good team for the implementation. Yah, in life, some things may be modular. In computing sciences too. And more, when the engineering team finished the work, they need to wait for the implementation team start and finish the implementation. It without say about the documentation. Why I need two teams to do the job if I can do all with one team and in less time? Yes, why do we need two teams to do two jobs of complexities 1 & 1, while we would need one single team twice bigger to do a single job of complexity exp(1+1) in exp(1+1)^2 /(1+1) more time? (i have no pointer to give about this patatheory, but i think it could be relevant to answer to your need. BTW, it is an open and free one with no copyright, to relieve your budget). It's a dilemma of how to reduce the financial cost for projects. Not really. Just think about manpower sinkholes. And MathML it's a answer. :-D Could you give more details? I do not see how it could be so. AFAIK, this language does not include any formal interpreter nor formal to numerical converter. Does it? Samuel ___ dev mailing list dev@lists.scilab.org http://lists.scilab.org/mailman/listinfo/dev
Re: [Scilab-Dev] Q/A about features
Hello Samuel, it's good to see you have appreciated the suggestions.
I will put the suggestions in the wiki, but before I want to discuss
the ideas here to enrich and mature.
I have not answered just below your words because my email client is
not working well. :-P
> > * Salvar o ambiente
(Save the environment)
Yes, in the client I can do it.But how to do in the CLI in text
mode?It's to automate many operations and to expand the work in text
mode.
> > * Interaction with users
Yes, including I have reported issues in bugzilla.But I don't know if a
user takes more than 2 minutes to be able to report something, I
believe he can desist.
> > * % keyword as comment:
The need it's simple.
Not always a software can do everything.
So to improve the job sometimes it needs to switch to a software to
other in differents steps of the job.
Think about the differences in modules, hardware consumption, and cost.
In common they always will have the same scripts. So better
compatibility make this exchange smoother.
For example:
I can convert images to HDF5 with Octave.
Run a personalized module in SciLAB
Make the heavy job in a cluster with MatLAB.
And Finish in SciLAB with the personalized module.
And no always what macro will run in each software it's clear. So it's
necessary that the 3 share the same scripts.
> > * Colors and arrays with names instead of numbers
Is about to reduce the number of steps and gain productivity. Including
for more complex selections.
Using the %nan i need:1. to write the selection operation2. make a
array copy 3. put %nan in the selection4. display the array5. restore
the original array
It's 5 steps. :-O If the selection it's always the same so why do not
reduce the visualization to a only single step ?
Repeat this 100 times on the same day it's a nightmare.
And more, the idea to put colors and words can be used to introduce
sense in the numbers.A matrix may have not only one word, but many.
For example: Normal, Caution, Danger, Error.The colors can have the
same meaning.
I don't know If I'm being clear
> > * Pool of formulas
It's a dilemma.
It I have a good engineering team but they can not or do not know how
to make a good SciLAB implementation, I will need another good team for
the implementation.And more, when the engineering team finished the
work, they need to wait for the implementation team start and finish
the implementation.
It without say about the documentation.
Why I need two teams to do the job if I can do all with one team and in
less time?
It's a dilemma of how to reduce the financial cost for projects.And
MathML it's a answer. :-D
-- Amanda Osvaldo
On Thu, 2017-02-23 at 00:03 +0100, Samuel Gougeon wrote:
> Hello Amanda,
>
>
>
> Thanks for this very long list of nice ideas and/or questions.
>
> I guess you already know this wiki page:
> https://wiki.scilab.org/GSoC_project_proposal
>
> You may add some ideas and open new pages to describe them.
>
> But thinking about is quite easy. Doing things is nicer :)
>
>
>
> I would like just answer to some of your questions:
>
>
>
> > * Salvar o ambiente
> >
> > I have not found in
> > the manual, is there any way to save all the variables and
> > macros
> > that are in memory in an HDF5 file to then load it again?
> >
> >
> > This allows a person
> > to stop what he is doing and continue later from where he
> > left
> > off.
> >
> >
>
>
>
> This is possible from the console menu File => Save/Load
> environment.."
>
> This includes opened figures, that are also saved and
> restored.
>
>
>
>
> >
> >
> >
> >
> > @page { margin: 2cm }
> > h1 { margin-bottom: 0.21cm }
> > h1.western { font-family: "Liberation Sans", sans-
> > serif; font-size: 18pt }
> > h1.cjk { font-family: "Droid Sans Fallback"; font-size:
> > 18pt }
> > h1.ctl { font-family: "Droid Sans Devanagari"; font-
> > size: 18pt }
> > p { margin-bottom: 0.25cm; line-height: 120% }
> > a:link { so-language: zxx }
> >
>
>
> > *
> > Interaction with users
> >
> > I
> > believe it's important to give more openness to users.
> >
> > Starting
> > with a button to allow reporting a bug from SciLAB itself
> > and
> > send an
> > evaluation about usability.
> > Maybe
> > even the documentation should be open for users to re-
> > write
> >
> >
> >
>
>
>
> Each embedded help page has a "Report a bug" link available at
> its
> top, pointing to the Bugzilla online bug tracker. The only
> issue
> with this link is that it appears also for pages of external
> modules, whose bugs trackers are not on bugzilla.
>
>
