GSoC project.

[Apekshit Sharma]

Some parts of this message have been removed. Learn more about Nabble's security policy.

Hi all

i want to be a part of open source development ,and it will be good to start with Google SoC.(as i will get a mentor :) :) ).here is my proposal for the same .i would request respected developers to please give a look at my proposal and make me aware of my flaws (hope there are not too many of them).please suggest the changes i should do in it to make it better.

please ignore randomness in bold characters.i tried to put them correct,but alas..actually i typed it all in openoffice word.

Thanks in advance.

PROPOSAL

Hi all

i want to be a part of open source development ,and it will be good to start with Google SoC.(as i will get a mentor :) :) ).here is my proposal for the same .i would request respected developers to please give a look at my proposal and make me aware of my flaws (hope there are not too many of them).please suggest the changes i should do in it to make it better.

Thanks in advance.

PROPOSAL

Name:Apekshit Sharma

 

Email Address:apeksharma@...

 

Freenode IRC Nick:apeksharma

 

Location (City, Country and/or Time Zone):Jaipur, rajasthan , INDIA. .+5:30 hours (i do my programming job in night mostly )

 

Proposal Name:I have decided to put my efforts in development of Kalzium.I searched through a number of KDEEDU project but i choose kalzium. My choice has three reasons behind it which have mentioned in motivation section.

Here is my experience of last two days :

“In past two days, i have tried out kalzium a number of times and had gone through a part of source code.During this , i came up with lots of interesting ideas which i think if are implemented in it, will not only add a lot of new features to it but will also increase its scope in the sense that it will surve to larger no of people.”

So here are my ideas,(and which i wish kalzium should have).I have also assigned priority to the tasks as major or minor. This proiority is based on my judgement (what i feel) of increment, a particular task will bring to kalzium as a whole:

1)(major)Drawback:Currently in chemical equation solver, only those ionic equations can be solved in which all of the RHS and LHS ions are given and that also when coffecients of sufficient no(depends on eqn)

of ions are given .This limits the use of this feature to a large extent as in high schools ,there are a no of ionic eqn's to be solved in which only the medium is given(acidic or basic) and ions(H+ and OH- ) change accordingly.This was only an example which limits ionic equation solver from its full capacity.

Addition:I will implement a seperate solver for ionic equations which will incorporate the above feature of a) selecting a medium of reaction and b)just writing the ions of LHS and RHS it will solve the eqaution without need of any cofficients (as it currently does).

2)(major)Right now in kalzium ,user can only calculate only molecular mass of a compound and the tab for the same is not so appealing.I would like to add to it features like

a)given amount of substance(weight),calculate moles

b)given weight of substance added and volume of solution/solvent,calculate molarity/molality.

c)calculate density ,strengths of solutions given weight of compound and volume of solvent and vice-versa.

d)giving results on mixture of two solutions.

And many other such small things which would provide kalzium with wider range of functions and serve more diverse problems of users.


3)(major)I would also like to add a feature which i'd like to call “compound finder”, and will do the fllowing:

once clicking on this feature ,the user when clicks on an element ,it will show all the compounds of that element .On subsquint selection of elements,it will keep on filtering compounds common to all selected elements before and the new one.Also i will provide additional filters like , setting min or/and max no. of atoms of each element.in the compound,giving max or/and min molecular mass, or like comp. should belong to a certain family (like alcohols,aldehydes,proteins,etc) .etc

It will be a lot useful as it will provide functionality of finding compounds by filters.

4)(minor)Current timeline tab is very premitive and not very charming.I would like to make major modifications to it like using openGL to build an more interective subwindow which will show periodic table building from scratch and showing time progressing.

Also i can add a feature , of selecting two years and showing elements discovered between these two years.

Motivation for Proposal / Goal:As mentioned before, I had three major motivations behind taking up this project which are:

1. am very keen of learning new things and i am at a stage ,at which ,i think i should try out different things and choose which is best suited for me.And open source programming is a vast feild and seem to be very intersting to me, i am inclined towards giving it a shot.Who knows, this might be a start to long journey ahead.

2. I am India National Chemistry Olympiad Gold Medalitst (only 35 all over india got it :) ).So , i have clear understanding of all the concepts i intend to implement .So like any other contender ,i won't have to devote any extra time in understanding those concepts (Instead i will use it learn open GL andQt )

3. I made a program in C in my 10^th class which implemented the 2^nd proposal above to a large extent.And so i am pretty confident that “i will do it”.

Implementation Details:

Proposal 1:I will be using the already built parser to fetch the data,and then use it in my algo.

It wouldn't be a tough nut to crack designing the algo, as i already know how are the things done in paper, and with my programming exp. (6 years is enough right now i guess :) ), i know it could be done with no pain.

Proposal 2:Its the same as above.Designing the better algo (as i have already mentioned, i did a part of it in my 10^th project ie 3 years before) and then coding it into c++.

proposal 3:It will require some more knowledge – Qt , which i plan to acquire in advance before 23^rd may.Also it will need designing a new format of storing data (which maps necessary data with the compound and organizes data in an efficient manner of which i have a crude picture in mind). What will be a challenge is,how much efficient can i make it ,as it will be dealing with database of lakhs of compound (and evergrowing database).

Proposal 4: For its implementation ,i will be learning OpenGl during the same time as Qt.Although i know i bit of opengl (to be strict ,i saw a part of opengl programming for this project and it was really easy understanding it),and i am very much confident that it can be done easily.

Tentative Timeline:My semester exams will be over by 21^st april ,so my GSoC project will start from 22^nd april

22^nd april-23^rd may (official start) – learning OpenGl,Qt and Open source programming style.

Now when the SoC starts officially:

90 days left....

15 days – implementation of the (better) ionic equation solver.(1^st proposal)

75 days left....

10 days – implementation of additional aspects to molecular mass calculator (2^nd proposal)

65 days left ....

20 days – implementation of “compound finder” (3^rd proposal). (i have planned it late because i feel it to be really a major feature and thus will need more experience which i will gain from 25 days of working)

45 days left...

5 days – implementation of timeline.(4^th proposal)

40 days left...

next 5-10 days – reviewing things

and the remaining time i plan to discuss as many things as i can, with my mentor, which can be implemented to make kalzium better.I'll try to add maybe 1, or 2 more things to kalzium which will be like “compound finder”.

I intend to work for about 8 hours daily (as i am completely free during my summers owing to no pre-engagements).This (7x 8) hours/week would certainly include posting my weekly achievements on to kde development list and discussing with mentor about what better can i do.

Do you have other obligations from late May to early August (school, work, etc.)?:I have no obligations during my summers, nor have i set any large goals .So i feel it will be only GsoC and me during summer.Infact , I will be free by 21^st april as my semster ends the same day. So , I will be having a good 32 days in hand before the start of GsoC (it think it will only be formal to say that it starts on 23^rd May,infact programming will start well before it) to learn OpenGL and KDE coding style .

 

About Me (let us know who you are!):

I am Apekshit Sharma.I am currently pursuing for my B.Tech degree and am a  first year student in Computer Science and Engineering Dept.at Indian Institute of Technology,Bombay(IIT,Bombay).(www.iitb.ac.in).

I have been doing programming for 6 years.

I started my programming in C,

Currently i know- C,C++,Scheme and PHP.

_______________________________________________
Kalzium mailing list
Kalzium@...
https://mail.kde.org/mailman/listinfo/kalzium

[Carsten Niehaus-2]

Am Sunday 29 March 2009 00:19:20 schrieb Apekshit Sharma:
> *1)(**major)**Drawback:Currently in chemical equation solver, only those
> ionic equations can be solved in which all of the RHS and LHS ions are
> given and that also when coffecients of sufficient no(depends on eqn)*

What is RHS and LHS?

> of ions are given .This limits the use of this feature to a large extent as
> in high schools ,there are a no of ionic eqn's to be solved in which only
> the medium is given(acidic or basic) and ions(H+ and OH- ) change
> accordingly.This was only an example which limits ionic equation solver
> from its full capacity.
>
>
>
>  Addition:I will implement a seperate solver for ionic equations which will
> incorporate the above feature of a) selecting a medium of reaction and
> b)just writing the ions of LHS and RHS it will solve the eqaution without
> need of any cofficients (as it currently does).

How will you implement it. Benoît Jacob (principle author of Eigen) told me a
year or two ago, that Eigen2 can be used for the math behind this. Right now
the code is written in OCaml which nobody on this list is able to maintain. I
would really like to get rid of the OCaml Stuff as Thomas Nagy (the author of
that code) won't improve it.

Carsten

--
Fatal Error: Found [MS-Windows] System -> Repartitioning Disk for Linux...
(By cbbrown@..., Christopher Browne)


_______________________________________________
Kalzium mailing list
Kalzium@...
https://mail.kde.org/mailman/listinfo/kalzium

[Carsten Niehaus-2]

Am Sunday 29 March 2009 00:19:20 schrieb Apekshit Sharma:
> 3)*(major)*I would also like to add a feature which i'd like to call
> “compound finder”, and will do the fllowing:
>
> once clicking on this feature ,the user when clicks on an element ,it will
> show all the compounds of that element .

All compounds? That would mean clicking on Carbon will show you a billion
carbonic compounds (ethanol, methane, butane, dodecane, dedecanol, ...)? That
is impossible. You will need a database for that. But hey, that database
exists :-)

http://chem-file.sourceforge.net/data/index_en.html

Look for example at

http://chem-file.sourceforge.net/data/alkanes/2_3-dimethylbutane_en.html

That could indeed result in some kind of frontend for chem-file. I am CC'ing
Jerome, the author of that database. Perhaps he has some comments on this.

Jerome?

> On subsquint selection of
> elements,it will keep on filtering compounds common to all selected
> elements before and the new one.Also i will provide additional filters like
> , setting min or/and max no. of atoms of each element.in the
> compound,giving max or/and min molecular mass, or like comp. should belong
> to a certain family (like alcohols,aldehydes,proteins,etc) .etc
>
> It will be a lot useful as it will provide functionality of finding
> compounds by filters.

Yes, I very much agree with this. This could even mean improving the
Strigi-Chemical project from 2007!

http://api.kde.org/playground-api/utils-apidocs/strigi-chemical/html/index.html

For comments on Strigi chemical I am CC'ing Egon Willighagen. He knows a lot
about Strigi and chemistry :-)

Egon?

Carsten

--
Microsoft is not the answer.
Microsoft is the question.
NO (or Linux) is the answer.
(Taken from a .signature from someone from the UK, source unknown)


_______________________________________________
Kalzium mailing list
Kalzium@...
https://mail.kde.org/mailman/listinfo/kalzium

[Egon Willighagen-5]

On Sun, Mar 29, 2009 at 1:07 PM, Carsten Niehaus <cniehaus@...> wrote:

> Am Sunday 29 March 2009 00:19:20 schrieb Apekshit Sharma:
>> 3)*(major)*I would also like to add a feature which i'd like to call
>> “compound finder”, and will do the fllowing:
>>
>> once clicking on this feature ,the user when clicks on an element ,it will
>> show all the compounds of that element .
>
> All compounds? That would mean clicking on Carbon will show you a billion
> carbonic compounds (ethanol, methane, butane, dodecane, dedecanol, ...)? That
> is impossible. You will need a database for that. But hey, that database
> exists :-)

Alternatively, I can imagine that it finds all compounds which only
contain that single element... so, for carbon it would find graphite,
buckyball, diamond.

>> On subsquint selection of
>> elements,it will keep on filtering compounds common to all selected
>> elements before and the new one.

That's quite impossible indeed... and I suggest linking to compound
databases on the internet for this, and open the search results in
Konqueror instead.

Jerome's database will only do organic compounds...

>> Also i will provide additional filters like
>> , setting min or/and max no. of atoms of each element.in the
>> compound,giving max or/and min molecular mass, or like comp. should belong
>> to a certain family (like alcohols,aldehydes,proteins,etc) .etc
>>
>> It will be a lot useful as it will provide functionality of finding
>> compounds by filters.
>
> Yes, I very much agree with this. This could even mean improving the
> Strigi-Chemical project from 2007!
>
> http://api.kde.org/playground-api/utils-apidocs/strigi-chemical/html/index.html
>
> For comments on Strigi chemical I am CC'ing Egon Willighagen. He knows a lot
> about Strigi and chemistry :-)

Strigi comes into play, particularly, when looking for compounds on
the local hard disk, that match the search criteria.

cc-ing, Alexandr.

Egon

--
Post-doc @ Uppsala University
http://chem-bla-ics.blogspot.com/
_______________________________________________
Kalzium mailing list
Kalzium@...
https://mail.kde.org/mailman/listinfo/kalzium

[Apekshit Sharma]

On Sun, Mar 29, 2009 at 4:31 PM, Carsten Niehaus <cniehaus@...> wrote:

Am Sunday 29 March 2009 00:19:20 schrieb Apekshit Sharma:
> *1)(**major)**Drawback:Currently in chemical equation solver, only those

> ionic equations can be solved in which all of the RHS and LHS ions are

> given and that also when coffecients of sufficient no(depends on eqn)*

What is RHS and LHS?

RHS is right hand side of equation
LHS -left hand side

> of ions are given .This limits the use of this feature to a large extent as
> in high schools ,there are a no of ionic eqn's to be solved in which only
> the medium is given(acidic or basic) and ions(H+ and OH- ) change
> accordingly.This was only an example which limits ionic equation solver
> from its full capacity.
>
>
>
>  Addition:I will implement a seperate solver for ionic equations which will
> incorporate the above feature of a) selecting a medium of reaction and
> b)just writing the ions of LHS and RHS it will solve the eqaution without
> need of any cofficients (as it currently does).

How will you implement it. Benoît Jacob (principle author of Eigen) told me a
year or two ago, that Eigen2 can be used for the math behind this. Right now
the code is written in OCaml which nobody on this list is able to maintain. I
would really like to get rid of the OCaml Stuff as Thomas Nagy (the author of
that code) won't improve it.

well ,
the way i planned to do it didn't took into account Eigen2 or anything like this .infact, i wasn't aware of anything like this.
now that you have told me there is this thing "Eigen 2 " it might simplify things for me .i might be able to do this task well before time.
to be honest, i haven't heard of OCaml before .although now i have read about what it is.And if it can help me simplify things, i can give it a shot .afterall , i will also learn something new.

 

Carsten

--
Fatal Error: Found [MS-Windows] System -> Repartitioning Disk for Linux...
(By cbbrown@..., Christopher Browne)

_______________________________________________
Kalzium mailing list
Kalzium@...
https://mail.kde.org/mailman/listinfo/kalzium

[Alexandr Goncearenco-2]

On Sun, 2009-03-29 at 13:13 +0200, Egon Willighagen wrote:

> On Sun, Mar 29, 2009 at 1:07 PM, Carsten Niehaus <cniehaus@...> wrote:
> > Am Sunday 29 March 2009 00:19:20 schrieb Apekshit Sharma:
> >> 3)*(major)*I would also like to add a feature which i'd like to call
> >> “compound finder”, and will do the fllowing:
> >>
> >> once clicking on this feature ,the user when clicks on an element ,it will
> >> show all the compounds of that element .
> >
> > All compounds? That would mean clicking on Carbon will show you a billion
> > carbonic compounds (ethanol, methane, butane, dodecane, dedecanol, ...)? That
> > is impossible. You will need a database for that. But hey, that database
> > exists :-)
>
> Alternatively, I can imagine that it finds all compounds which only
> contain that single element... so, for carbon it would find graphite,
> buckyball, diamond.

This sounds ok for carbon, but I can't imagine such a query for
non-carbon single element compounds.
Compound filter sounds like a more vial feature than element filter,
though I would love to be able to get a list of metal-containing
proteins (PDB format) on my local disk by metal element, for example.

> >> On subsquint selection of
> >> elements,it will keep on filtering compounds common to all selected
> >> elements before and the new one.
>
> That's quite impossible indeed... and I suggest linking to compound
> databases on the internet for this, and open the search results in
> Konqueror instead.
>
> Jerome's database will only do organic compounds...
>
> >> Also i will provide additional filters like
> >> , setting min or/and max no. of atoms of each element.in the
> >> compound,giving max or/and min molecular mass, or like comp. should belong
> >> to a certain family (like alcohols,aldehydes,proteins,etc) .etc
> >>
> >> It will be a lot useful as it will provide functionality of finding
> >> compounds by filters.
> >
> > Yes, I very much agree with this. This could even mean improving the
> > Strigi-Chemical project from 2007!
> >
> > http://api.kde.org/playground-api/utils-apidocs/strigi-chemical/html/index.html
> >
> > For comments on Strigi chemical I am CC'ing Egon Willighagen. He knows a lot
> > about Strigi and chemistry :-)
>
> Strigi comes into play, particularly, when looking for compounds on
> the local hard disk, that match the search criteria.

This is a great possibility to integrate Strigi with Kalzium. Although
strigi-chemical demands a separate search/query frontend ideally, having
a simple client embedded in Kalzium might be the best start.

Apekshit, are you planning to submit all three proposals? You have
marked them all as "(major)", so it's quite hard to see which one you'd
prefer most.

I've been a GSoC 2007 student working on strigi-chemical metadata
extraction. Current status is that you can perform searches based on
these chemical metadata, but we lack a frontend. Whether or not this
should be part of Kalzium is up to Carsten, but I assure that making a
GUI to search for chemical compounds on you harddrive (which have
already been indexed and are there in the strigi database) is an
exciting and rewarding project. This will also revive strigi-chemical,
as it is hardly used by anyone by this time, mainly due to the missing
GUI.

Apekshit, check my blog-report: http://neksa.blogspot.com/  and the
source: http://websvn.kde.org/trunk/playground/utils/strigi-chemical/
Should you have any questions, don't hesitate to ask.

Best wishes,
Alex.

_______________________________________________
Kalzium mailing list
Kalzium@...
https://mail.kde.org/mailman/listinfo/kalzium

[Apekshit Sharma]

On Mon, Mar 30, 2009 at 7:06 PM, Alexandr Goncearenco <neksa@...> wrote:

On Sun, 2009-03-29 at 13:13 +0200, Egon Willighagen wrote:
> On Sun, Mar 29, 2009 at 1:07 PM, Carsten Niehaus <cniehaus@...> wrote:
> > Am Sunday 29 March 2009 00:19:20 schrieb Apekshit Sharma:
> >> 3)*(major)*I would also like to add a feature which i'd like to call
> >> “compound finder”, and will do the fllowing:
> >>
> >> once clicking on this feature ,the user when clicks on an element ,it will
> >> show all the compounds of that element .
> >
> > All compounds? That would mean clicking on Carbon will show you a billion
> > carbonic compounds (ethanol, methane, butane, dodecane, dedecanol, ...)? That
> > is impossible. You will need a database for that. But hey, that database
> > exists :-)
>
> Alternatively, I can imagine that it finds all compounds which only
> contain that single element... so, for carbon it would find graphite,
> buckyball, diamond.

This sounds ok for carbon, but I can't imagine such a query for
non-carbon single element compounds.
Compound filter sounds like a more vial feature than element filter,
though I would love to be able to get a list of metal-containing
proteins (PDB format) on my local disk by metal element, for example.

this gives me a good idea.
when i meet my current proposal goal, i will be able to help you uptill -

"I would love to be able to get a list of metal-containing

proteins (PDB format) on my local disk"(only till here)

as when you'll select a metal , it will give you all compounds containing that metal from the database.and i plan to show the results grouped by families(which will obviously contain protein family) or something(not yet decided other grouping criteria).
But

 "by metal element"

has given  me  idea of incorporating a feature which might do something like this:
save the result.(which a currently plan to do)
take a part of result
and then regroup it on some other basis like by metal (obviously it will lead to same compound being in multiple groups),non-metal,etc
in your case, it turns out to be be metal.
although i won't take it as a part of my mainstream project(sorry), as i would be concentrating mainly on my first 3 proposal. but if i get time , i'll be glad to incorporate this feature. :-)

> >> On subsquint selection of
> >> elements,it will keep on filtering compounds common to all selected
> >> elements before and the new one.
>
> That's quite impossible indeed... and I suggest linking to compound
> databases on the internet for this, and open the search results in
> Konqueror instead.
>
> Jerome's database will only do organic compounds...
>
> >> Also i will provide additional filters like
> >> , setting min or/and max no. of atoms of each element.in the
> >> compound,giving max or/and min molecular mass, or like comp. should belong
> >> to a certain family (like alcohols,aldehydes,proteins,etc) .etc
> >>
> >> It will be a lot useful as it will provide functionality of finding
> >> compounds by filters.
> >
> > Yes, I very much agree with this. This could even mean improving the
> > Strigi-Chemical project from 2007!
> >
> > http://api.kde.org/playground-api/utils-apidocs/strigi-chemical/html/index.html
> >
> > For comments on Strigi chemical I am CC'ing Egon Willighagen. He knows a lot
> > about Strigi and chemistry :-)
>
> Strigi comes into play, particularly, when looking for compounds on
> the local hard disk, that match the search criteria.

This is a great possibility to integrate Strigi with Kalzium. Although
strigi-chemical demands a separate search/query frontend ideally, having
a simple client embedded in Kalzium might be the best start.

Apekshit, are you planning to submit all three proposals? You have
marked them all as "(major)", so it's quite hard to see which one you'd
prefer most.

I've been a GSoC 2007 student working on strigi-chemical metadata
extraction. Current status is that you can perform searches based on
these chemical metadata, but we lack a frontend. Whether or not this
should be part of Kalzium is up to Carsten, but I assure that making a
GUI to search for chemical compounds on you harddrive (which have
already been indexed and are there in the strigi database) is an
exciting and rewarding project. This will also revive strigi-chemical,
as it is hardly used by anyone by this time, mainly due to the missing
GUI.

Apekshit, check my blog-report: http://neksa.blogspot.com/  and the
source: http://websvn.kde.org/trunk/playground/utils/strigi-chemical/
Should you have any questions, don't hesitate to ask.

thanks a lot alex,
i have given a look at your blog, but i won't be getting much time before 1st april owing to my approaching lab exam :-( 
It does serve my purpose of what i intend to do.
So i plan to have a better look at the same on 1st and i'll be surely giving you pain by asking questions :-)
thank you once again for helping me out.

Best wishes,
Alex.

to all
it's really a great experience discussing the things on mail and exchanging ideas with people whom i didn't even know, not even two days before.
this open source stuff is turning out to be a lot interesting.
i didn't know that there was so much help available and that there were people always ready to help and give suggestions.
it's like, i have got a lot things needed ,but i know , there more waiting for me.
thanks a lot to all.

apekshit sharma

_______________________________________________
Kalzium mailing list
Kalzium@...
https://mail.kde.org/mailman/listinfo/kalzium

[Carsten Niehaus-2]

Moin moin

> This is a great possibility to integrate Strigi with Kalzium. Although
> strigi-chemical demands a separate search/query frontend ideally, having
> a simple client embedded in Kalzium might be the best start.
>
> Apekshit, are you planning to submit all three proposals? You have
> marked them all as "(major)", so it's quite hard to see which one you'd
> prefer most.
>
> I've been a GSoC 2007 student working on strigi-chemical metadata
> extraction. Current status is that you can perform searches based on
> these chemical metadata, but we lack a frontend. Whether or not this
> should be part of Kalzium is up to Carsten,

While I think that Kalzium is not the ideal place for this it is a good
starting point. Why not host the code inside Kalzium and implementing it
as a QWidget which is loaded from inside Kalzium. Later on we could then
reuse that code in apps like Avogadro or Plasma-applet.

> but I assure that making a
> GUI to search for chemical compounds on you harddrive (which have
> already been indexed and are there in the strigi database) is an
> exciting and rewarding project. This will also revive strigi-chemical,
> as it is hardly used by anyone by this time, mainly due to the missing
> GUI.

Full Ack. And an important one!

Carste
_______________________________________________
Kalzium mailing list
Kalzium@...
https://mail.kde.org/mailman/listinfo/kalzium