www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - Some standard protocols in Phobos

reply bearophile <bearophileHUGS lycos.com> writes:
I think it will eventually become important to have some standard protocols in
Phobos, like one for images like this:
http://www.python.org/dev/peps/pep-0368/

Do you like?

and similar ones for:
- matrices
- basic 2D pixel-oriented graphics.
- basic database usage

Bye,
bearophile
Dec 12 2011
next sibling parent Somedude <lovelydear mailmetrash.com> writes:
Le 12/12/2011 13:55, bearophile a écrit :
 I think it will eventually become important to have some standard protocols in
Phobos, like one for images like this:
 http://www.python.org/dev/peps/pep-0368/
 
 Do you like?
 
 and similar ones for:
 - matrices
 - basic 2D pixel-oriented graphics.
 - basic database usage
 
 Bye,
 bearophile

I don't see the need for this.
Dec 12 2011
prev sibling next sibling parent =?utf-8?Q?Simen_Kj=C3=A6r=C3=A5s?= <simen.kjaras gmail.com> writes:
On Mon, 12 Dec 2011 13:55:53 +0100, bearophile <bearophileHUGS lycos.com>  
wrote:

 I think it will eventually become important to have some standard  
 protocols in Phobos, like one for images like this:
 http://www.python.org/dev/peps/pep-0368/

 Do you like?

 and similar ones for:
 - matrices
 - basic 2D pixel-oriented graphics.
 - basic database usage

These would be nice, but are by no means a necessity.
Dec 12 2011
prev sibling parent reply bcs <bcs example.com> writes:
On 12/12/2011 04:55 AM, bearophile wrote:
 I think it will eventually become important to have some standard protocols in
Phobos, like one for images like this:
 http://www.python.org/dev/peps/pep-0368/

 Do you like?

 and similar ones for:
 - matrices
 - basic 2D pixel-oriented graphics.
 - basic database usage

 Bye,
 bearophile

Likely the easiest way to ram that in in would be to pick one, design an interface, implement 3-5 concrete examples and submit a pull request. I.e. create a de-facto standard. I think there is already one in the works for DBs .
Dec 12 2011
next sibling parent reply Jonathan M Davis <jmdavisProg gmx.com> writes:
On Monday, December 12, 2011 20:07:37 bcs wrote:
 On 12/12/2011 04:55 AM, bearophile wrote:
 I think it will eventually become important to have some standard
 protocols in Phobos, like one for images like this:
 http://www.python.org/dev/peps/pep-0368/
 
 Do you like?
 
 and similar ones for:
 - matrices
 - basic 2D pixel-oriented graphics.
 - basic database usage
 
 Bye,
 bearophile

Likely the easiest way to ram that in in would be to pick one, design an interface, implement 3-5 concrete examples and submit a pull request. I.e. create a de-facto standard. I think there is already one in the works for DBs .

A pull request would not be sufficient. Any major additions to Phobos must go through a review process on the newsgroup and be voted into Phobos. Pull requests are generally for smaller additions and bug fixes rather than whole new modules and the like. So, unless you're talking about just a function or two here, then this kind of stuff is likely to require a formal review. That being said, the main way to get this kind of stuff into Phobos is to just buckle down and do it. If you're the first to come up with a new piece of functionality and are willing to work it through the review process, then you can get stuff added that way. If you sit around, hoping that someone else will do it, it may never happen. - Jonathan M Davis
Dec 12 2011
parent bcs <bcs example.com> writes:
On 12/12/2011 08:17 PM, Jonathan M Davis wrote:
 On Monday, December 12, 2011 20:07:37 bcs wrote:
 On 12/12/2011 04:55 AM, bearophile wrote:
 I think it will eventually become important to have some standard
 protocols in Phobos, like one for images like this:
 http://www.python.org/dev/peps/pep-0368/

 Do you like?

 and similar ones for:
 - matrices
 - basic 2D pixel-oriented graphics.
 - basic database usage

 Bye,
 bearophile

Likely the easiest way to ram that in in would be to pick one, design an interface, implement 3-5 concrete examples and submit a pull request. I.e. create a de-facto standard. I think there is already one in the works for DBs .

A pull request would not be sufficient. Any major additions to Phobos must go through a review process on the newsgroup and be voted into Phobos. Pull requests are generally for smaller additions and bug fixes rather than whole new modules and the like. So, unless you're talking about just a function or two here, then this kind of stuff is likely to require a formal review.

OK, so maybe I over simplified a bit and omitted to mention some of the later steps. :b
 That
 being said, the main way to get this kind of stuff into Phobos is to just
 buckle down and do it. If you're the first to come up with a new piece of
 functionality and are willing to work it through the review process, then you
 can get stuff added that way. If you sit around, hoping that someone else will
 do it, it may never happen.

Let me clarify what I was trying to say: What he said. :o)
Dec 12 2011
prev sibling next sibling parent reply Brad Roberts <braddr puremagic.com> writes:
On 12/12/2011 8:46 PM, bcs wrote:
 Let me clarify what I was trying to say: What he said. :o)

Let me rephrase your clarification. :) The best way to get something done is to do it. Throwing out an idea and hoping someone else does it for you is fairly ineffective.
Dec 12 2011
parent reply bearophile <bearophileHUGS lycos.com> writes:
Brad Roberts:

 The best way to get something done is to do it.

I think the point of view about your opinion differers across different nations of the world. That POV leads to results like: http://delight.sourceforge.net/ Before doing things some people want to discuss if it's a good idea, and to know if others are interested in its results/consequences. Bye, bearophile
Dec 12 2011
parent bcs <bcs example.com> writes:
On 12/12/2011 11:13 PM, bearophile wrote:
 Brad Roberts:

 The best way to get something done is to do it.

I think the point of view about your opinion differers across different nations of the world. That POV leads to results like: http://delight.sourceforge.net/ Before doing things some people want to discuss if it's a good idea, and to know if others are interested in its results/consequences.

True, but that will happen any way before a pull request gets blessed. And if the package doesn't pass muster, you can still publish it somewhere else.
Dec 13 2011
prev sibling parent reply Jesse Phillips <jessekphillips+d gmail.com> writes:
On Mon, 12 Dec 2011 20:07:37 -0800, bcs wrote:

 Likely the easiest way to ram that in in would be to pick one, design an
 interface, implement 3-5 concrete examples and submit a pull request.
 I.e. create a de-facto standard.
 
 I think there is already one in the works for DBs .

Haha, it is good to know your audience. I was going to make a similar statement, but saw bearophile had made the post. Know is involvement in the community I realized he wasn't asking to get these in, but as he says poll for interest and opinions. For such I don't really have much to comment. The only one of those I can really understand what it would mean to have a standard protocol for is the DB. The others, I'm not sure what it is solving that isn't addressed by actually having it in Phobos (Why not have matrices and not just their protocol?).
Dec 13 2011
next sibling parent bcs <bcs example.com> writes:
On 12/13/2011 06:46 PM, Jesse Phillips wrote:
 On Mon, 12 Dec 2011 20:07:37 -0800, bcs wrote:

 Likely the easiest way to ram that in in would be to pick one, design an
 interface, implement 3-5 concrete examples and submit a pull request.
 I.e. create a de-facto standard.

 I think there is already one in the works for DBs .

Haha, it is good to know your audience. I was going to make a similar statement, but saw bearophile had made the post. Know is involvement in the community I realized he wasn't asking to get these in, but as he says poll for interest and opinions. For such I don't really have much to comment. The only one of those I can really understand what it would mean to have a standard protocol for is the DB. The others, I'm not sure what it is solving that isn't addressed by actually having it in Phobos (Why not have matrices and not just their protocol?).

"Sequential" data is to ranges as matrices are to what?
Dec 13 2011
prev sibling parent bearophile <bearophileHUGS lycos.com> writes:
Jesse Phillips:

 The others, I'm not sure what it is solving that isn't addressed 
 by actually having it in Phobos (Why not have matrices and not just their 
 protocol?).

The focus of my post was mostly on the protocol for images. The problem with having actual code in Phobos is that a good matrix library, or a good image library, is a lot of well written code. And it can't be good enough for everyone, so other people will want to invent other image or matrix libraries. The point of having a protocol in Phobos is to avoid giving too much work to Phobos devs, and at the same time have a standard API to use a matrix or image. Bye, bearophile
Dec 14 2011