Re: [Glsdk-devel] [GLm] api

[H. H. <hen...@gm...>]

Mul and div for colors: I have the idea of implementing colors as different
types altogether:

typedef ... GLMcol3f;
typedef ... GLMcol4f;

glmColSet3f
glmColAdd3f
...

So in this respect, mul and div for vectors are redundant.

About glmVecScale: yes that is already in the current API.

About doubles: yes, doubles are not yet supported in the GL (afaik), but
they will be probably included when the hardware supports them. So we can
prepare for this and use the -f prefix. Also, for colors we might want to
have things like -ub also. So including the type postfix makes sense.

We also had an earlier short discussion in #glsdk if we should have:

glmVectorSet2f
glmMatrixSet2f
glmMatrixIdentity2f

instead of

glmVecSet2f
glmMatSet2f
...

But we decided to keep the short ones. Also this is open for discussion, as
is everything dealing with the GLM API. Comments are most welcome.


On Mon, Aug 25, 2008 at 11:53 PM, Mars_999 <ma...@si...> wrote:

>  Oops I should have said Mul, and Div aren't defined for vectors. Just
> clearing up my posting.
>
> ----- Original Message -----
> *From:* Mars_999 <ma...@si...>
> *To:* gls...@li...
> *Sent:* Monday, August 25, 2008 3:44 PM
> *Subject:* Re: [Glsdk-devel] [GLm] api
>
> glmVecSet2f
> glmVecAdd2f
> glmVecLength2f
> glmVecTransform2f
> glmMatSet2f
> glmMatIdentity2f
>
> I vote for this.
>
> Correct, in mathematics vectors aren't defined, but I was thinking ahead
> for colors or other uses, having these would be nice, but anyone who is
> coding graphics should understand this already?
>
> Also don't forget to put in a glmVecScale2f? I guess we could put double in
> but, IIRC double was going to get dropped? And we aren't doing OpenCL or
> CUDA here AFAIK, so double would kind of wasteful in terms of time spent.
>
> ----- Original Message -----
> *From:* Henri Häkkinen <hen...@gm...>
> *To:* gls...@li...
> *Sent:* Monday, August 25, 2008 3:17 PM
> *Subject:* Re: [Glsdk-devel] [GLm] api
>
> We are currently using floats, but may want to reserve the possibility of
> havings doubles too. Therefore we use the -f postfix. This is also the GL
> convention of doing things:
>
> glVertex2f
>
> and therefore we have:
>
> glmVecSet2f
>
> instead of:
>
> glmVec2Set
>
> Also, we don't have glmVecDiv or glmVecMul. Why? Because vector division or
> multiplication are not defined in linear-algebra. There are two kinds of
> vector products: cross and dot. These are handled by the -Cross and -Dot
> entry points. Member-wise div or mul are not mathemtically defined
> operations in vector calculus AFAIK.
>
> We have now several open issues with the GLM API:
>
> 1) Should we provide -Mul and -Div for vectors?
> My opinion: no, as they are not defined in mathematics and they generally
> are not needed
>
> 2) Should we use short naming convention or is the current convention fine?
>
> The current convention is, for example:
>
> glmVecSet2f
> glmVecAdd2f
> glmVecLength2f
> glmVecTransform2f
> glmMatSet2f
> glmMatIdentity2f
>
> The proposed short convention would be, for example:
>
> glmVecLen2f
> glmVecTrans2f
> glmMatIdent2f
> glmMatDet2f
>
> 3) Should we drop the -f postfix, or should we keep it for future possible
> double types?,
>
> glmVecSet2f (the current API) vs. glmVecSet2 (the proposed)
>
> My opinion: we should keep the -f, it is more OpenGL-like and we reserve
> the possibility of -d
>
> 4) Should we have the dimension of the type in the end of the entry-point
> (like in OpenGL) or in the middle?
>
> glmVecSet2f (the current API) vs. glmVec2Setf (the proposed)
>
> My opinion: glmVecSet2f, it is more OpenGL-like
>
>
> Issues are open for discussion. Please comment.
>
>
>
> On Mon, Aug 25, 2008 at 10:10 PM, Mars_999 <ma...@si...> wrote:
>
>>  I haven't looked at the lib yet, but aren't we just going to work with
>> floats all the time?
>>
>> glmVec2Sub
>> glmVec3Sub
>> glmVec4Sub
>>
>> glmVec2Add
>> glmVec3Add
>> glmVec4Add
>>
>> glmVec2Mul
>> glmVec2Div
>>
>> glmVec2Norm
>> glmVec2Transf
>> glmVec3Cross
>> glmVec3Dot
>>
>> I like that format...
>>
>>   ----- Original Message -----
>> *From:* Henri Häkkinen <hen...@gm...>
>> *To:* glsdk-devel <gls...@li...>
>> *Sent:* Monday, August 25, 2008 2:04 PM
>> *Subject:* [Glsdk-devel] [GLm] api
>>
>> The GLM API seems at this moment to use two sets of naming styles. Compare
>> these two for example:
>>
>> glmVecSub2f
>> glmVecLength2f
>>
>> The first function is using a short form for the operation ('sub' as from
>> 'substract') while the other is using long form. Should we simplify the API
>> to look like this:
>>
>> glmVecSub2f
>> glmVecLen2f
>>
>> Althought on the other hand, how do we shorten such stuff as Normalize,
>> Transform, Identity, Inverse? This does not necessary look good:
>>
>> glmVecNorm
>> glmVecTrans
>> glmMatTrans
>> glmMatInv
>> glmMatIdent
>>
>> Is this even a big issue? Comments?
>>
>> If we decide to keep the API as it is now, I think we should lock it in
>> and begin to outline the API reference for it. For later versions, I would
>> like to implement such stuff as colors, quaternions and maybe planes and
>> matrix stacks but this can wait. Bigger priority is to get the current
>> version working and tested for.
>>
>> For Korval: I outlined some API ref in a plain text format. You may find
>> it in 'docs/sdk_reference/glm-apiref' in the latest revision. You may use
>> this as ideas for the DocBook API ref and this file needs to get removed
>> once you migrate it into DocBook.
>>
>> --
>> Henri 'henux' Häkkinen
>>
>>  ------------------------------
>>
>> -------------------------------------------------------------------------
>> This SF.Net email is sponsored by the Moblin Your Move Developer's
>> challenge
>> Build the coolest Linux based applications with Moblin SDK & win great
>> prizes
>> Grand prize is a trip for two to an Open Source event anywhere in the
>> world
>> http://moblin-contest.org/redirect.php?banner_id=100&url=/
>>
>> ------------------------------
>>
>> _______________________________________________
>> Glsdk-devel mailing list
>> Gls...@li...
>> https://lists.sourceforge.net/lists/listinfo/glsdk-devel
>>
>>
>> -------------------------------------------------------------------------
>> This SF.Net email is sponsored by the Moblin Your Move Developer's
>> challenge
>> Build the coolest Linux based applications with Moblin SDK & win great
>> prizes
>> Grand prize is a trip for two to an Open Source event anywhere in the
>> world
>> http://moblin-contest.org/redirect.php?banner_id=100&url=/
>> _______________________________________________
>> Glsdk-devel mailing list
>> Gls...@li...
>> https://lists.sourceforge.net/lists/listinfo/glsdk-devel
>>
>>
>
>
> --
> Henri 'henux' Häkkinen
>
>  ------------------------------
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's
> challenge
> Build the coolest Linux based applications with Moblin SDK & win great
> prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
>
> ------------------------------
>
> _______________________________________________
> Glsdk-devel mailing list
> Gls...@li...
> https://lists.sourceforge.net/lists/listinfo/glsdk-devel
>
>  ------------------------------
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's
> challenge
> Build the coolest Linux based applications with Moblin SDK & win great
> prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
>
> ------------------------------
>
> _______________________________________________
> Glsdk-devel mailing list
> Gls...@li...
> https://lists.sourceforge.net/lists/listinfo/glsdk-devel
>
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's
> challenge
> Build the coolest Linux based applications with Moblin SDK & win great
> prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> Glsdk-devel mailing list
> Gls...@li...
> https://lists.sourceforge.net/lists/listinfo/glsdk-devel
>
>


-- 
Henri 'henux' Häkkinen