Fwd: Squeak 5 on Raspberry Pi
Locked
12
12
Re: Fwd: Squeak 5 on Raspberry Pi
 
|
Hi Stef,
On Fri, Mar 25, 2016 at 2:27 PM, stepharo <[hidden email]> wrote:
yes, of course :-)
+1
I don't think this is either or. I don't think Hilaire is saying "don't do feature-rich releases". I think Hilaire is saying "do separate feature-rich and consolidation releases". I think Hilaire is making a good suggestion of having some releases being for new functionality and some for consolidation. So perhaps the community could schedule a Pharo N, followed by a Pharo N.1, or schedule a Pharo M, where M is even (or odd) and a Pharo N, where N is odd (or even), and have the N.1, or N is odd (or even) releases being consolidation releases as Hilaire describes, with no new features and only bug fixes and performance improvements. If the community can manage it, it could do one new feature, followed by one consolidation release a year. But if so, the community needs to be serious about the consolidation releases and put real effort into them. The advantage for the broader user base is clear; there is a steady stream of releases that more conservative users can use, that exhibits good stability and better performance than the bleeding edge. Also, there's no reason why these release cycles can't overlap. The only time they must not overlap is when the community needs to focus on the new release. So for example, for two months of the year, six months apart, the community should focus on the new release, be it consolidation or feature-rich, and make sure we release promptly and correctly. But the other ten months of the year there's no reason why the community cannot work on either release. This requires infrastructure such as two repositories, one for each, such as pharo-features and pharo-stable, and the discipline to separate one's work correctly, but it could be a really good thing. I'm sure others can think this through better than I. What do others suggest?
_,,,^..^,,,_ best, Eliot |
|
|
2016-03-25 19:32 GMT-03:00 Eliot Miranda <[hidden email]>:
In my opinion, that sounds like a very good and more serious approach for the whole community. The only problem I see is an explosion of Pharo flavors like happened in Squeak years ago, because many people want promotion of their own image (which is ok but also reflects lack of consensus), but is better than having new core projects with developers abandoning before time or obscure decisions imposing a new debugger without ANY poll. Now I don't want to waste time on definitions of "solid", but I can tell you Pharo now feels slow. About having
to ask "spend some time optimising tools" just a few weeks before a release... Sorry guys I know you work a lot on making good quality stuff, but seems like the current way of integrating tools **is NOT working**. Where is the policy for integrating packages and tools into the core? Where are the priorities? Why they are priorities? How many votes gets each?Want more contributors? Start with clean visible rules and listen to users. We all want UFFI, Xtreams, mini-image, new UI frameworks & builder, etc. but not at any cost. Hernán |
Re: Fwd: Squeak 5 on Raspberry Pi
|
|
I totally do not agree with many of the arguments in this thread.
But I do not want to discuss them under this label. Please start a new thread, if you want to. > On 26 Mar 2016, at 05:40, Hernán Morales Durand <[hidden email]> wrote: > > > > 2016-03-25 19:32 GMT-03:00 Eliot Miranda <[hidden email]>: > Hi Stef, > > On Fri, Mar 25, 2016 at 2:27 PM, stepharo <[hidden email]> wrote: > So guys, you do not want Xtreams (and prefer to use Streams that have been "designed" decades ago) ;D ? > no bootstrapped core (clement you do not want to have a mini image :) ? and a new UI frameworks? > > yes, of course :-) > > I personnally want to have new widgets, a real UI builder and massively cleaning Spec. > Now I would like to have multiple tool sets - I understand that people like the new debugger (I do not like it) - > I want the possibility to have a mini tools tool set. > > +1 > > If you want to clean Pharo you can start cleaning Komitter stupid use of state pattern generating > a lot of garbage instead of having a single animated morph. > > We should clean Versionner- I have the impression that half of the classes are not mandatory. > > I don't think this is either or. I don't think Hilaire is saying "don't do feature-rich releases". I think Hilaire is saying "do separate feature-rich and consolidation releases". > > I think Hilaire is making a good suggestion of having some releases being for new functionality and some for consolidation. So perhaps the community could schedule a Pharo N, followed by a Pharo N.1, or schedule a Pharo M, where M is even (or odd) and a Pharo N, where N is odd (or even), and have the N.1, or N is odd (or even) releases being consolidation releases as Hilaire describes, with no new features and only bug fixes and performance improvements. If the community can manage it, it could do one new feature, followed by one consolidation release a year. But if so, the community needs to be serious about the consolidation releases and put real effort into them. > > The advantage for the broader user base is clear; there is a steady stream of releases that more conservative users can use, that exhibits good stability and better performance than the bleeding edge. > > Also, there's no reason why these release cycles can't overlap. The only time they must not overlap is when the community needs to focus on the new release. So for example, for two months of the year, six months apart, the community should focus on the new release, be it consolidation or feature-rich, and make sure we release promptly and correctly. But the other ten months of the year there's no reason why the community cannot work on either release. This requires infrastructure such as two repositories, one for each, such as pharo-features and pharo-stable, and the discipline to separate one's work correctly, but it could be a really good thing. I'm sure others can think this through better than I. What do others suggest? > > In my opinion, that sounds like a very good and more serious approach for the whole community. The only problem I see is an explosion of Pharo flavors like happened in Squeak years ago, because many people want promotion of their own image (which is ok but also reflects lack of consensus), but is better than having new core projects with developers abandoning before time or obscure decisions imposing a new debugger without ANY poll. > > Now I don't want to waste time on definitions of "solid", but I can tell you Pharo now feels slow. > > About having to ask "spend some time optimising tools" just a few weeks before a release... Sorry guys I know you work a lot on making good quality stuff, but seems like the current way of integrating tools **is NOT working**. > > Where is the policy for integrating packages and tools into the core? > Where are the priorities? Why they are priorities? How many votes gets each? > > Want more contributors? > Start with clean visible rules and listen to users. > > We all want UFFI, Xtreams, mini-image, new UI frameworks & builder, etc. but not at any cost. > > Hernán |
|
|
In reply to this post by Nicolai Hess-3-2
me too. Now you will not my core. Because there is no UI no announcement...
No I cannot let you to say that. I'm sorry. We spent 4 months cleaning Nautilus and Rubric is not optimal but we decided that forcing to adapt to Rubric was a good move for the next one. We did it to help - breakpoints - QA feedback Rubric was pushed in Pharo over the summer. So if we cannot change something as important that that more than 8 months before the release then we should better stop to do pharo. Because for Rubric ***I*** planned it in advance with a stabilisation phase. Now Pharo 50 got far too many new features We should have not include Spur and release Pharo 50 without it and keep new FFI and Spur for Pharo 60 and keep GTTools for Pharo 60 If this is your analysis then we are ok. Now you cannot tell me that pushing rubric was a mess. The state of the system (in particular the lack of good widgets) is a problem. Look at the keybindings why the keybinding is still the mess: simply because all the widgets and tools just nicely harcoded them. It does not mean that we should not fix them but you cannot fight day long against spur migration bugs fight day long against FFI glitches fix integration bugs of GT and get more steam for the rest Now let me tell you frankly I prefer to build useless mini new languages than fighting with ugly widgets so why did I supervised and worked with a guy to improve and push rubric? Because it was needed. indeed it is badly documented. Now it has examples. Now synectique and moose have been using in their product rubric. No alain helped each time we asked him but he has severe family problems (the kind of problems that you do not want to have but you have to face). What can we say? I have no idea why igor disappeared and kind of divorced. Now the design of TxText is nice and we will have to invest. Bricks already used it and people looking at it mentioned that it is good. So the objectives is to drop rubric (because it is a hack and we know it but a hack which supports embellisment and icons) and use TxText. But you see you are becoming the most aware for athens and what we should do is document document. Now I got ****EXHAUSTED**** to document things that I did not build or use daily. This effort for me is gigantic and I cannot do it each time. Do you think that roassal is extended privately Athens. I would be surprised. I know that we all like what you did with the widgets because blocers can work in athens with default widgets and our goal is to throw all the morphic layer away and only use Athens. Now we should give feedback to blocers
We need a UI Builder and spec is a way to build widgets. Now before we get the perfect solution we need to make sure that we clean it. I would like to do that with Peter. But again there is no magic: some part of spec are ugly because of design but others are ugly because the widgets are poor. Now for me I have no problem cleaning Spec even if at the end we replace it by something else. We did that for the compiler, we are doing it for Morphic, ....
I know nicolai and I understand your frustration and I understand it:) I thank you everyday for that. I think that we should remove things from Pharo So the most important point for us is to get in place a process so that we can avoid to get monolithic again For example we need a process to have the possibility to remove project from the image and still build and modify an image with them. It will not change the problem that when a bug is there we have a bug but it should lower the stress. Where? Nautilus is much better to me. I used Versionner and it is working. CodeCritics I got some glitches with refactorings Do you have some issues? I thought that it was not the case and I do not think so. So this is side effect of the cleaning of foundations. The problems is that we cannot block people working on the bootstrap forever. So nicolai what I would do is a roadmap for Pharo 60 - there will be no Spur and FFI :) - so it will be consolidation - I would like to have release every six months (but it should be discussed) - for me I would like to have - cleaning Spec - cleaning another time nautilus - cleaning versionner - cleaning Komitter - Now we have epicea waiting - So Xtreams will be probably for later. |
Re: Fwd: Squeak 5 on Raspberry Pi
|
|
In reply to this post by stepharo
Yes, I want all of the good stuff. Let's always remember that when the platoon soldiers stop complaining, it is time to watch your back for a knife. Looks like we are a long way from there. Phil On Fri, Mar 25, 2016 at 10:27 PM, stepharo <[hidden email]> wrote:
|
Re: Fwd: Squeak 5 on Raspberry Pi
|
|
In reply to this post by Eliot Miranda-2
In Tiki, our versioning policy is like this: Going on nicely since 2009. Phil On Fri, Mar 25, 2016 at 11:32 PM, Eliot Miranda <[hidden email]> wrote:
|
Re: Fwd: Squeak 5 on Raspberry Pi
|
|
In reply to this post by stepharo
On Sat, Mar 26, 2016 at 9:19 AM, stepharo <[hidden email]> wrote:
This is just too much indeed. But we are learning a lot by doing this. Phil |
Re: Fwd: Squeak 5 on Raspberry Pi
|
|
In reply to this post by stepharo
Le 26/03/2016 09:19, stepharo a écrit :
> > But you see you are becoming the most aware for athens > and what we should do is document document. > Now I got ****EXHAUSTED**** to document things that I did not build or > use daily. This effort for me is gigantic > and I cannot do it each time. In my opinion Pharo should refuse to include any class without at least a one line documentation for the small classes. And when someone review a bug correction and see that an important class have a really small documentation it should be a stop for the integration. I never saw a Java class without documentation on oracle site. This will slow down the integration of some project, but since everyone agree that pharo included too many project, this should not be a problem. At least we would integrate only projects understandable by the others. -- Cyril Ferlicot http://www.synectique.eu 165 Avenue Bretagne Lille 59000 France |
signature.asc (836 bytes)
Re: Fwd: Squeak 5 on Raspberry Pi
|
|
In reply to this post by philippeback
Hi Phil,
|
|
|
In reply to this post by EstebanLM
Le 25/03/2016 13:31, Esteban Lorenzano a écrit :
> Hi, > > actually there is a tree implementation based on fast table, but I’m not > sure is production ready (Cyril can say something about). > But anyway since I never understood why we have a tree there (since we > will “open” attribute in next panel), I do agree, strongly :) Hi, There is a working FTTree with examples and doc in Pharo 5 and there is a Glamour adaptor as for FastTable (With examples in GLMExampleBrowser). Since it is not use for now they might be some bugs. I think that the code would benefits some review from experienced people since I am still far from being as good as most of the developers here :) For example some things I do not like: - Since I warped data into Items object I had to add a "realElementAt:" method in the dataSource. This is not wanted in my opinion but I don't know how to remove it and I don't have the time to work on it anymore. - Some features cause infinite loop if the number of children is infinite (for example, search in all the items). I wrote it in the documentation when that might happen but maybe we can reduce the number of feature that can provoke that? - I think that the search/filtering time can be improve when we have to check in the complete tree. I would like that experienced people check the code before we really use it because if there is some change in the design to do, we should do it now. > not sure, I open and it takes some seconds to actually show anything. I > thought it was caused because of the catalog refresh to populate > projects, but I’m not sure. > In any case, load catalog projects in spotter is cool, but we need to > find a better implementation. One that does not freezes spotter until it > refresh (remember, Pharo is also used in places with crapy connection or > not connection at all). > > yes! > > if you want to suffer, do this experiment: > 1. open a Pharo2 and open debugger on anything > 2. then Pharo3, 4 and 5… feel the incremental slowdown > 3. cry :’( > > Esteban > > Cyril Ferlicot http://www.synectique.eu 165 Avenue Bretagne Lille 59000 France |
Re: Fwd: Squeak 5 on Raspberry Pi
|
|
In reply to this post by CyrilFerlicot
Good luck with things like Roassal or Seaside. As of Java, sure, but Javadoc is only a small part. Try to understand Swing with Javadocs and no thick book, good luck with that. The Java Tutorial is another matter. Phil On Sat, Mar 26, 2016 at 3:50 PM, Cyril Ferlicot D. <[hidden email]> wrote: Le 26/03/2016 09:19, stepharo a écrit : |
Re: Fwd: Squeak 5 on Raspberry Pi
|
|
In reply to this post by Eliot Miranda-2
My point was about the Versioning approach. Tiki is about all kinds of things. That's what you get with millions of lines of PHP code and, indeed, millions of downloads etc. It is a community building tool. Check https://tiki.org/Features and mix and match them as most things can be combined orthogonally (like security meet calendar, or trackers meet articles). That's a hell of a powerful tool I can tell you. I used it to support several businesses and it rocks. Like Squeak and Pharo, it may lead to some "what's this thing?" impressions. Phil On Sat, Mar 26, 2016 at 3:59 PM, Eliot Miranda <[hidden email]> wrote:
|
|
|
In reply to this post by CyrilFerlicot
Le 26/3/16 15:50, Cyril Ferlicot D. a écrit : > Le 26/03/2016 09:19, stepharo a écrit : >> But you see you are becoming the most aware for athens >> and what we should do is document document. >> Now I got ****EXHAUSTED**** to document things that I did not build or >> use daily. This effort for me is gigantic >> and I cannot do it each time. > In my opinion Pharo should refuse to include any class without at least > a one line documentation for the small classes. And when someone review > a bug correction and see that an important class have a really small > documentation it should be a stop for the integration. > > I never saw a Java class without documentation on oracle site. at least you made my day comparing pharo with oracle I felt proud :) > > This will slow down the integration of some project, but since everyone > agree that pharo included too many project, this should not be a > problem. At least we would integrate only projects understandable by the > others. |
|
|
In reply to this post by CyrilFerlicot
Hi,
> On Mar 26, 2016, at 5:04 PM, Cyril Ferlicot D. <[hidden email]> wrote: > > Le 25/03/2016 13:31, Esteban Lorenzano a écrit : >> Hi, >> >> actually there is a tree implementation based on fast table, but I’m not >> sure is production ready (Cyril can say something about). >> But anyway since I never understood why we have a tree there (since we >> will “open” attribute in next panel), I do agree, strongly :) > > Hi, > > There is a working FTTree with examples and doc in Pharo 5 and there is > a Glamour adaptor as for FastTable (With examples in GLMExampleBrowser). > Since it is not use for now they might be some bugs. I think that the > code would benefits some review from experienced people since I am still > far from being as good as most of the developers here :) > For example some things I do not like: > - Since I warped data into Items object I had to add a "realElementAt:" > method in the dataSource. This is not wanted in my opinion but I don't > know how to remove it and I don't have the time to work on it anymore. > - Some features cause infinite loop if the number of children is > infinite (for example, search in all the items). I wrote it in the > documentation when that might happen but maybe we can reduce the number > of feature that can provoke that? > - I think that the search/filtering time can be improve when we have to > check in the complete tree. > > I would like that experienced people check the code before we really use > it because if there is some change in the design to do, we should do it > now. As far as I saw, this is just a tree, not a tree with table (like the one used in the Raw presentation). Am I wrong? Cheers, Doru >> not sure, I open and it takes some seconds to actually show anything. I >> thought it was caused because of the catalog refresh to populate >> projects, but I’m not sure. >> In any case, load catalog projects in spotter is cool, but we need to >> find a better implementation. One that does not freezes spotter until it >> refresh (remember, Pharo is also used in places with crapy connection or >> not connection at all). >> >> yes! >> >> if you want to suffer, do this experiment: >> 1. open a Pharo2 and open debugger on anything >> 2. then Pharo3, 4 and 5… feel the incremental slowdown >> 3. cry :’( >> >> Esteban >> >> > > -- > Cyril Ferlicot > > http://www.synectique.eu > > 165 Avenue Bretagne > Lille 59000 France > -- www.tudorgirba.com www.feenk.com "Be rather willing to give than demanding to get." |
«
Return to Pharo Smalltalk Developers
|
1 view|%1 views
| Free forum by Nabble | Edit this page |