So, here's the deal of why I wanted to delete objects. My first idea was actually to create a dynamic number of outlets, but that didn't work, cause connections got lost. So I thought, well, maybe have a maximum number of outlets and delete them, but that turned out to be very much uglier than I thought and I haven't felt like trying yet.
But now I see there's this [initbang] external just for that... too bad that, like I said, I cannot afford to have an external dependency... And then I've also seen there was a discussion of including [initbang] functionality in Pd Vanilla, which I believe it's of vital importance if we want Pd Vanilla to have any real functionality of providing powerful externals.
I usually see many people here promoting this idea of "doing it yourself with plainly vanilla", well, seems that many times that's really just hard or impossible, so if we really want that, the [initbang] functionality would be imperative. And, as long as we're at it, an easy message to pd for deleting objects would be nice too.
So, I'm talking about what I found here: https://lists.puredata.info/pipermail/pd-list/2016-01/112868.html
And I ask, what came out of that? Is the jury still out?
Thanks
2017-10-18 11:22 GMT-02:00 IOhannes m zmoelnig zmoelnig@iem.at:
On 2017-10-18 15:15, Alexandre Torres Porres wrote:
2017-10-18 3:35 GMT-02:00 IOhannes m zmölnig zmoelnig@iem.at:
there is a reason, why a library like iemguts has an object like [canvasdelete]: you cannot do this otherwise.
But is there any particular reason not to have a simple message to Pd
that
would delete an object? So one could be "in luck" and not have to rely on externals or painful in the back hacks?
iirc, there was a feature-request (cum implementation) on the sf tracker and it was turned down.
fgmasdr IOhannes
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/ listinfo/pd-list
On 2017-10-19 06:19, Alexandre Torres Porres wrote:
And I ask, what came out of that? Is the jury still out?
it has been implemented exactly as miller suggested:
Pd will not include [initbang], but externals can (now) implement it. iemguts' already has an [initbang] that you can use.
fgmasdr IOhannes
I see... somehow I got terribly confused and thought there was a discussion to include it in vanilla, which would make sense to me. Maybe it was the conversation about meeting a "proper design" of initbang/closebang that led me to think that was on the table. Sorry for the confusion.
Well, I guess this is now turning into a request to bring such functionalities to vanilla. My point being that it'd be of vital importance if we want Pd Vanilla to have any real functionality of providing ways to program powerful external abstractions.
If not, since I do not want dependancies to other external libraries, the only way would be to "steal" it for my library, huh? What are the ethics on that?
cheers
2017-10-19 5:43 GMT-02:00 IOhannes m zmoelnig zmoelnig@iem.at:
On 2017-10-19 06:19, Alexandre Torres Porres wrote:
And I ask, what came out of that? Is the jury still out?
it has been implemented exactly as miller suggested:
Pd will not include [initbang], but externals can (now) implement it. iemguts' already has an [initbang] that you can use.
fgmasdr IOhannes
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/ listinfo/pd-list
On 10/19/2017 03:20 PM, Alexandre Torres Porres wrote:
Well, I guess this is now turning into a request to bring such functionalities to vanilla. My point being that it'd be of vital importance if we want Pd Vanilla to have any real functionality of providing ways to program powerful external abstractions.
yawn. sorry, but you are 10 years late. we've had this discussion. multiple times. the conclusion was that Pd now provides the infrastucture to implement [initbang] as an external. personally, i think this is good enough, and most likely all we will ever get. i'm pretty tired of issues being re-raised over more than a decade. (it obviously show that there is interest in the problem. but we now do have a solution and your point currently is mainly that you don't like the particula implementation of it)
If not, since I do not want dependancies to other external libraries, the only way would be to "steal" it for my library, huh? What are the ethics on that?
(you probably deserve eternal damnation for such a thing.) most likely you will discover another issue that has been neatly solved in iemguts in a week or two. and steal that. and then you discover that iemguts hsa other nifty things. and just steal all of them, until you have stolen it all. just for the sake of "not having a dependency".
apart from that, iemguts is licensed under the GPL-2+. everybody is free to use it under this license (but note, that e.g. ELSE is currently not compatible with GPL)
gfmadsr IOhannes
On 19/10/17 16:23, IOhannes m zmölnig wrote:
On 10/19/2017 03:20 PM, Alexandre Torres Porres wrote:
If not, since I do not want dependancies to other external libraries,
Why?
the only way would be to "steal" it for my library, huh? What are the ethics on that?
(you probably deserve eternal damnation for such a thing.) most likely you will discover another issue that has been neatly solved in iemguts in a week or two. and steal that. and then you discover that iemguts hsa other nifty things. and just steal all of them, until you have stolen it all. just for the sake of "not having a dependency".
And then original upstream iemguts gets some new bugfixes, and your copy is still buggy.
apart from that, iemguts is licensed under the GPL-2+. everybody is free to use it under this license (but note, that e.g. ELSE is currently not compatible with GPL)
Important point.
Claude
2017-10-19 13:58 GMT-02:00 Claude Heiland-Allen claude@mathr.co.uk:
On 19/10/17 16:23, IOhannes m zmölnig wrote:
On 10/19/2017 03:20 PM, Alexandre Torres Porres wrote:
If not, since I do not want dependancies to other external libraries,
Why?
Why not? I mean, sorry, but I'm not comfortable or happy getting into a kind of discussion on where I have to defend personal preferences and ideas that I don't really care to defend if others disagree. But if you have considerations against it, I'm happy to hear, though I reserve myself the right to make up my own mind.
thanks
I'm not sure if there's much to "steal" in the case of [initbang] (I don't like the word BTW. in the open source world you shouldn't have to "steal" things). but I'm wondering: how can one reimplement [initbang] and publish it under a different, more permissive license? basically it's just an API call and a version check... there's not much to do differently. shall [initbang] therefore always be GPL?
Gesendet: Donnerstag, 19. Oktober 2017 um 17:23 Uhr Von: "IOhannes m zmölnig" zmoelnig@iem.at An: pd-list@lists.iem.at Betreff: Re: [PD] initbang in vanilla (was Re: how to destroy/delete an object with dynamic patching?)
On 10/19/2017 03:20 PM, Alexandre Torres Porres wrote:
Well, I guess this is now turning into a request to bring such functionalities to vanilla. My point being that it'd be of vital importance if we want Pd Vanilla to have any real functionality of providing ways to program powerful external abstractions.
yawn. sorry, but you are 10 years late. we've had this discussion. multiple times. the conclusion was that Pd now provides the infrastucture to implement [initbang] as an external. personally, i think this is good enough, and most likely all we will ever get. i'm pretty tired of issues being re-raised over more than a decade. (it obviously show that there is interest in the problem. but we now do have a solution and your point currently is mainly that you don't like the particula implementation of it)
If not, since I do not want dependancies to other external libraries, the only way would be to "steal" it for my library, huh? What are the ethics on that?
(you probably deserve eternal damnation for such a thing.) most likely you will discover another issue that has been neatly solved in iemguts in a week or two. and steal that. and then you discover that iemguts hsa other nifty things. and just steal all of them, until you have stolen it all. just for the sake of "not having a dependency".
apart from that, iemguts is licensed under the GPL-2+. everybody is free to use it under this license (but note, that e.g. ELSE is currently not compatible with GPL)
gfmadsr IOhannes
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list
2017-10-19 13:23 GMT-02:00 IOhannes m zmölnig zmoelnig@iem.at:
your point currently is mainly that you don't like the particular implementation of it)
Not sure if I get what you mean by me not liking the "particular implementation of it". I'd just say my point is that I believe this functionality is worth having in vanilla. Though I know now I'm late on this discussion and, apparently, it has been decided it should not have such functionality. Not to argue about it, but I'd appreciate if anyone would tell me the considerations behind this decision.
2017-10-19 12:45 GMT-02:00 Jonathan Wilkes jancsika@yahoo.com:
There were many requests to include it in Pd Vanilla. Check
the mailing list archives.
Or I'll do that and see what I can find if no one wants to tell me :)
thanks
On Don, 2017-10-19 at 17:23 +0200, IOhannes m zmölnig wrote:
On 10/19/2017 03:20 PM, Alexandre Torres Porres wrote:
Well, I guess this is now turning into a request to bring such functionalities to vanilla. My point being that it'd be of vital importance if we want Pd Vanilla to have any real functionality of providing ways to program powerful external abstractions.
i'm pretty tired of issues being re-raised over more than a decade.
Though I mostly agree with you, I disagree with the notion that raising the same thing several times is necessarily a bad thing. I'm living in a country where the people directly vote about some decisions. We would live in medieval times still - so to speak - if we hadn't voted about the same thing many times. 10 years later, the (Pd-) world might have changed a bit and suddenly implementing [initbang] in Pd-vanilla is considered nice and pretty.. How can you know?
(it obviously show that there is interest in the problem. but we now do have a solution
Good. :-)
apart from that, iemguts is licensed under the GPL-2+. everybody is free to use it under this license (but note, that e.g. ELSE is currently not compatible with GPL)
It seems as if the _license_ forces sane behavior. GPL forever!
Roman
2017-10-20 18:19 GMT-02:00 Roman Haefeli reduzent@gmail.com:
Though I mostly agree with you, I disagree with the notion that raising the same thing several times is necessarily a bad thing. I'm living in a country where the people directly vote about some decisions. We would live in medieval times still - so to speak - if we hadn't voted about the same thing many times. 10 years later, the (Pd-) world might have changed a bit and suddenly implementing [initbang] in Pd-vanilla is considered nice and pretty.. How can you know?
That was *exactly* what I was thinking, thank you for pointing that out so I can give you a +1
And yeah, 10 years ago was sorta ages ago for Pd, many things have happened, like the demise of Pd Extended. There's this libpd thing and the need for it to be more self sufficient. And this feels like such a basic core functionality that I really wonder why not to have it somehow. So I really wonder if the reasoning from 10 years ago is still as pertinent now for the community as whole.
cheers
ok, browsing the pd list archive for some data on "this has been discussed over and over", what I found was many discussions about initbang, making the external work in vanilla, things like that, but no real discussion on why not having it in vanilla.
I can see, from a couple of years ago, proposals for initbang in vanilla, but No actual development on the matter (namely, Alexandros Drymonitis and Dan Wilcox - source https://lists.puredata.info/pipermail/pd-list/2015-07/110826.html)
I also see a thread opened by jonathan wilkens in 2010, commenting on a patch to add initbang/colsebang to vanilla from 2006 (ID: 1544041), asking "what's the story on it?", saying that he's seen "*repeated hopes from various developers that these objects be included as internal objects in pd*."
- but this thread also had no feedback/conclusion - source
https://lists.puredata.info/pipermail/pd-list/2010-06/080219.html
Going back earlier, I see a message from 2008, where Matt Barber asks if initbang is on vanilla and IOhannes replies "no, and I'm not sure why", after that Matt Barbour "claims for its inclusion to, again, no response at all - source https://lists.puredata.info/pipermail/pd-list/2008-05/061952.html
And this, about 10 years ago, is when gave up digging deep through the mail archive trying to find a clear rejection why initbang should not be added to Vanilla. Ok, I might have missed something, but I did my best, and what I have makes me challenge this notion that this has been discussed and denied over and over. By the way, I have to say it's not the first time I hear something has been fully discussed on the mail list, but I actually do not find a clear closure on the matter...
I can ask again if someone had a source or fill me in on the veredict. I've also seen a same discussion come up, but when I respond to it, I give the source and what came out of that.
Don't get me wrong, I don't wanna reissue an argument that has been going on and on. And I don't wanna raise the same case once more and argue in favor of it when it has been turned out many times. It's just that I don't see it, can't find it. On the contrary, I found several requests in over many years that simply did not get real attention... but if there's in fact a clear and strong rejection, I don't really care, I won't mind, I'll find me another solution even if it involves making my own external - which is to say this is not a selfish motivation, I just think this could be an important request to be, as it really seems like a wish from many members of this community.
cheers
2017-10-20 21:17 GMT-02:00 Alexandre Torres Porres porres@gmail.com:
2017-10-20 18:19 GMT-02:00 Roman Haefeli reduzent@gmail.com:
Though I mostly agree with you, I disagree with the notion that raising the same thing several times is necessarily a bad thing. I'm living in a country where the people directly vote about some decisions. We would live in medieval times still - so to speak - if we hadn't voted about the same thing many times. 10 years later, the (Pd-) world might have changed a bit and suddenly implementing [initbang] in Pd-vanilla is considered nice and pretty.. How can you know?
That was *exactly* what I was thinking, thank you for pointing that out so I can give you a +1
And yeah, 10 years ago was sorta ages ago for Pd, many things have happened, like the demise of Pd Extended. There's this libpd thing and the need for it to be more self sufficient. And this feels like such a basic core functionality that I really wonder why not to have it somehow. So I really wonder if the reasoning from 10 years ago is still as pertinent now for the community as whole.
cheers
typo
I just think this could be an important request to be *reissued*,
2017-10-21 0:38 GMT-02:00 Alexandre Torres Porres porres@gmail.com:
ok, browsing the pd list archive for some data on "this has been discussed over and over", what I found was many discussions about initbang, making the external work in vanilla, things like that, but no real discussion on why not having it in vanilla.
I can see, from a couple of years ago, proposals for initbang in vanilla, but No actual development on the matter (namely, Alexandros Drymonitis and Dan Wilcox - source https://lists.puredata.info/pipermail/pd-list/2015-07/110826.html)
I also see a thread opened by jonathan wilkens in 2010, commenting on a patch to add initbang/colsebang to vanilla from 2006 (ID: 1544041), asking "what's the story on it?", saying that he's seen "*repeated hopes from various developers that these objects be included as internal objects in pd*." - but this thread also had no feedback/conclusion - source https://lists.puredata.info/pipermail/pd-list/2010-06/080219.html
Going back earlier, I see a message from 2008, where Matt Barber asks if initbang is on vanilla and IOhannes replies "no, and I'm not sure why", after that Matt Barbour "claims for its inclusion to, again, no response at all - source https://lists.puredata.info/pipermail/pd-list/2008-05/061952.html
And this, about 10 years ago, is when gave up digging deep through the mail archive trying to find a clear rejection why initbang should not be added to Vanilla. Ok, I might have missed something, but I did my best, and what I have makes me challenge this notion that this has been discussed and denied over and over. By the way, I have to say it's not the first time I hear something has been fully discussed on the mail list, but I actually do not find a clear closure on the matter...
I can ask again if someone had a source or fill me in on the veredict. I've also seen a same discussion come up, but when I respond to it, I give the source and what came out of that.
Don't get me wrong, I don't wanna reissue an argument that has been going on and on. And I don't wanna raise the same case once more and argue in favor of it when it has been turned out many times. It's just that I don't see it, can't find it. On the contrary, I found several requests in over many years that simply did not get real attention... but if there's in fact a clear and strong rejection, I don't really care, I won't mind, I'll find me another solution even if it involves making my own external - which is to say this is not a selfish motivation, I just think this could be an important request to be, as it really seems like a wish from many members of this community.
cheers
2017-10-20 21:17 GMT-02:00 Alexandre Torres Porres porres@gmail.com:
2017-10-20 18:19 GMT-02:00 Roman Haefeli reduzent@gmail.com:
Though I mostly agree with you, I disagree with the notion that raising the same thing several times is necessarily a bad thing. I'm living in a country where the people directly vote about some decisions. We would live in medieval times still - so to speak - if we hadn't voted about the same thing many times. 10 years later, the (Pd-) world might have changed a bit and suddenly implementing [initbang] in Pd-vanilla is considered nice and pretty.. How can you know?
That was *exactly* what I was thinking, thank you for pointing that out so I can give you a +1
And yeah, 10 years ago was sorta ages ago for Pd, many things have happened, like the demise of Pd Extended. There's this libpd thing and the need for it to be more self sufficient. And this feels like such a basic core functionality that I really wonder why not to have it somehow. So I really wonder if the reasoning from 10 years ago is still as pertinent now for the community as whole.
cheers
Ok then, cos i would also like vanilla initbang, and just to stir the hornets nest...
Why CAN'T it be done in vanilla?
On 10/20/2017 10:19 PM, Roman Haefeli wrote:
Though I mostly agree with you, I disagree with the notion that raising the same thing several times is necessarily a bad thing. I'm living in a country where the people directly vote about some decisions. We would live in medieval times still - so to speak - if we hadn't voted about the same thing many times. 10 years later, the (Pd-) world might have changed a bit and suddenly implementing [initbang] in Pd-vanilla is considered nice and pretty.. How can you know?
no sorry, that's not how it works.
while i totally agree on the general terms, this doesn't apply to the current issue.
the initbang discussion has *started* >10 years ago (for what it is worth: this is when I first implemented it). since then the issue was raised again and again. 1.5 years ago (see the post from miller that alex linked to) a decision was made, which was greeted with an "EVERYONE WINS" by the community (and i don't remember anybody objecting).
so what has changed in the last 20 months?
of course, we are living in a free world and you are free to raise your concerns about any decision, even it has reached consensus a minute ago. and if others don't want to get through all this again and again, they free to blacklist your address. though this is probably counterproductive on the long run.
gfadsr IOhannes
Ok, you have a good point and now I seem to finally understand what you mean.
It's not that this has been argued over and over only to get strong rejections as I was assuming. Seems I got confused yet again.
Apparently, the deal is that this has been long requested over and over, but all that ever came out of this was this possibility to have it as an external for Pd. Which is to say there was plenty awareness that people wanted it, but, nonetheless, a decision has been made by Miller and this is probably all we'll ever get. Am I close?
2017-10-21 5:08 GMT-02:00 IOhannes m zmölnig zmoelnig@iem.at:
1.5 years ago (see the post from miller that alex linked to) a decision was made, which was greeted with an "EVERYONE WINS" by the community (and i don't remember anybody objecting).
so what has changed in the last 20 months?
I don't think much has changed. But by that I mean a lot of people still seem to hope for that functionality in Vanilla and are still requesting for it. Yes making it possible for externals was obviously a well received step towards, unlikely subject to any objection by anybody, but I don't think that one or another person greeting "Yeah! Awesome" reflects the satisfaction of the whole community, or an agreement that actually having this in Vanilla from then on and forever would be crazy. Actually I can't see why there'd be much objection for not having this in Vanilla. Thus, still here we are asking for further solutions.
Ok, seems Miller has concerns. And I've already found my own solution (writting my own external), so I can let it go easily. Though I'm really curious and would like at least an explanation just to understand the issue.
2017-10-22 12:24 GMT-02:00 Dan Wilcox danomatika@gmail.com:
Maybe a further solution would be to have an optional [initbang] (or something similarly named) in the core "extra" externals which simply wraps the interface provided within vanilla and is overridable and/or optional when compiling.
I can think of an easier solution that doesn't even require another Pd object: [loadbang] could take a non zero argument to turn it into [initbang]. That'd be so easy to implement that even I could do it... (it's just action = LB_INIT instead of LB_LOAD). This would also be backwards compatible, of course. I could even go ahead and make a PR for that.
cheers
-
Alexandre Torres Porres -
Christof Ressi -
Claude Heiland-Allen -
IOhannes m zmoelnig -
IOhannes m zmölnig
-
Jonathan Wilkes -
Matt Davey -
Roman Haefeli