Re: [Pharo-project] Omnibrowser in 1.4
On Aug 29, 2011, at 7:54 PM, Lukas Renggli wrote:
No more for me.
Introducing more code into Pharo that depends on more parts of Pharo (RPackage, Announcement, Pragma, Ring, RB, Shout, ...) doesn't make it easier to maintain and change Pharo. Or did I misunderstand something about cohesion and coupling? :-)
With that philosophy, we can just declare Pharo as finished and do something else. My point of view is that I invent new abstractions *and than use them* on and for the system itself.
I disagree; I would like a small and stable Pharo in which crazy ideas can be realized. For that I don't need fancy abstractions, but a minimal, simple and absolutely stable system in which I can load and do whatever I want. Maybe this is just me?
Is the current system simple and minimal? Do you think the Pharo we have is good enough to have a future? Marcus -- Marcus Denker -- http://marcusdenker.de
No more for me.
Introducing more code into Pharo that depends on more parts of Pharo (RPackage, Announcement, Pragma, Ring, RB, Shout, ...) doesn't make it easier to maintain and change Pharo. Or did I misunderstand something about cohesion and coupling? :-)
With that philosophy, we can just declare Pharo as finished and do something else. My point of view is that I invent new abstractions *and than use them* on and for the system itself.
I disagree; I would like a small and stable Pharo in which crazy ideas can be realized. For that I don't need fancy abstractions, but a minimal, simple and absolutely stable system in which I can load and do whatever I want. Maybe this is just me?
Is the current system simple and minimal?
No, it is complex and it is getting bigger with every release.
Do you think the Pharo we have is good enough to have a future?
No, there is a lot to be improved. I think the future of Pharo is what can be built on top, not what can be integrated and forced upon everybody. Lukas -- Lukas Renggli www.lukas-renggli.ch
On 29 August 2011 19:47, Lukas Renggli <renggli@gmail.com> wrote:
No more for me.
Introducing more code into Pharo that depends on more parts of Pharo (RPackage, Announcement, Pragma, Ring, RB, Shout, ...) doesn't make it easier to maintain and change Pharo. Or did I misunderstand something about cohesion and coupling? :-)
With that philosophy, we can just declare Pharo as finished and do something else. My point of view is that I invent new abstractions *and than use them* on and for the system itself.
I disagree; I would like a small and stable Pharo in which crazy ideas can be realized. For that I don't need fancy abstractions, but a minimal, simple and absolutely stable system in which I can load and do whatever I want. Maybe this is just me?
Is the current system simple and minimal?
No, it is complex and it is getting bigger with every release.
No, i wouldn't say so. Most fixes and improvements are still about cleaning things out and fixing bugs. But not about new features. Even if you consider Zinc as a movement towards "getting bigger", consider an alternative: having library which follows standards, or keep using hacks which were not complete and useful only for most simple cases. I am also against growing system unless it is necessarily. In 1.3. i added a new 'non-interactive' mode and non-interactive ui manager. And this was necessarily, because we need a way to deal with "hanging images" in headless mode. Now we can run images on jenkins, knowing that it will never enter 'click ok to proceed' state.
Do you think the Pharo we have is good enough to have a future?
No, there is a lot to be improved. I think the future of Pharo is what can be built on top, not what can be integrated and forced upon everybody.
Lukas, there is nobody forcing anyone. How we (or anyone else) can force people to use Pharo? A new and 'unproven' tools are 'forced' into images mainly to collect feedback. And i think it is good strategy, because most people *including me* are too busy/lazy to go & download and install new stuff and try it out. I don't like some rough edges of Nautilus. And it is certainly having less features than OB. But.. I think if you compare it to basic browser which we have in our image.. you will see. (p.s. and yes i am totally agree with you - a sidebar buttons in Nautilus are ugly ;) -- Best regards, Igor Stasenko AKA sig.
Is the current system simple and minimal?
No, it is complex and it is getting bigger with every release.
No, i wouldn't say so.
Most fixes and improvements are still about cleaning things out and fixing bugs. But not about new features.
Yes, you are right. In fact Pharo 1.4 is roughly 1 MB smaller than Pharo 1.3.
Even if you consider Zinc as a movement towards "getting bigger", consider an alternative: having library which follows standards, or keep using hacks which were not complete and useful only for most simple cases.
Zinc is an excellent example, because it is fully backward compatible. I don't see that with RPackage, SystemAnnouncements, Ring, Shout (before Alan fixed it), with the proposed RB changes, ...
I am also against growing system unless it is necessarily. In 1.3. i added a new 'non-interactive' mode and non-interactive ui manager. And this was necessarily, because we need a way to deal with "hanging images" in headless mode. Now we can run images on jenkins, knowing that it will never enter 'click ok to proceed' state.
Yes, this is cool.
Do you think the Pharo we have is good enough to have a future?
No, there is a lot to be improved. I think the future of Pharo is what can be built on top, not what can be integrated and forced upon everybody.
Lukas, there is nobody forcing anyone. How we (or anyone else) can force people to use Pharo?
Right, but you can easily force people not to follow. I don't understand why all the people that still use Pharo 1.0 or 1.1 don't speak up?
A new and 'unproven' tools are 'forced' into images mainly to collect feedback. And i think it is good strategy, because most people *including me* are too busy/lazy to go & download and install new stuff and try it out.
Don't we have all these automatic builds with all these tools pre-loaded? There could be a list of links to these builds on the website and like/+1/report buttons. That would likely yield some discussions and give some time to investigate ... Lukas -- Lukas Renggli www.lukas-renggli.ch
On Aug 29, 2011, at 10:37 PM, Lukas Renggli wrote:
Is the current system simple and minimal?
No, it is complex and it is getting bigger with every release.
No, i wouldn't say so.
Most fixes and improvements are still about cleaning things out and fixing bugs. But not about new features.
Yes, you are right. In fact Pharo 1.4 is roughly 1 MB smaller than Pharo 1.3.
Even if you consider Zinc as a movement towards "getting bigger", consider an alternative: having library which follows standards, or keep using hacks which were not complete and useful only for most simple cases.
Zinc is an excellent example, because it is fully backward compatible. I don't see that with RPackage, SystemAnnouncements, Ring, Shout (before Alan fixed it), with the proposed RB changes, ...
lukas please do not ask impossible suff. This is boring after a while. I'm not doing pharo to be compatible with broken abstractions.
I am also against growing system unless it is necessarily. In 1.3. i added a new 'non-interactive' mode and non-interactive ui manager. And this was necessarily, because we need a way to deal with "hanging images" in headless mode. Now we can run images on jenkins, knowing that it will never enter 'click ok to proceed' state.
Yes, this is cool.
Do you think the Pharo we have is good enough to have a future?
No, there is a lot to be improved. I think the future of Pharo is what can be built on top, not what can be integrated and forced upon everybody.
Lukas, there is nobody forcing anyone. How we (or anyone else) can force people to use Pharo?
Right, but you can easily force people not to follow. I don't understand why all the people that still use Pharo 1.0 or 1.1 don't speak up?
A new and 'unproven' tools are 'forced' into images mainly to collect feedback. And i think it is good strategy, because most people *including me* are too busy/lazy to go & download and install new stuff and try it out.
Don't we have all these automatic builds with all these tools pre-loaded? There could be a list of links to these builds on the website and like/+1/report buttons. That would likely yield some discussions and give some time to investigate ...
Zinc is an excellent example, because it is fully backward compatible. I don't see that with RPackage, SystemAnnouncements, Ring, Shout (before Alan fixed it), with the proposed RB changes, ...
I do not understand how SystemAnnouncements could be fully backward compatible, then and this is the most important aspect of pharo. Same for RPackage, why should I reproduce broken semantics like overrides (it never worked over the last 5 years). if we do not fix the infrastructure to be aligned better stop and we will not stop so :) Now I will not comment anymore on this thread because I could have integrated 4 fixes but I spent time arguing the obvious instead. Stef
Hi guys, I am completely for inventing something new. However, I do not think this is the right model for doing so. In a sense we are getting static, we put some tools in the image and expect people to use them. I think that this is going to be contra productive. I believe that a much better model would be a dynamic one in which people can load the tools that make them more productive. Moreover, this model will force us, developers, to better target our tools. Make clear which are their objectives, showing their benefits, improve their design for understandability and extension and thus getting feedback and interested people. At the end better tools will emerge. In a word, better products, with a "natural selection" process. HTH, Jorge On Mon, Aug 29, 2011 at 10:37 PM, Lukas Renggli <renggli@gmail.com> wrote:
Is the current system simple and minimal?
No, it is complex and it is getting bigger with every release.
No, i wouldn't say so.
Most fixes and improvements are still about cleaning things out and fixing bugs. But not about new features.
Yes, you are right. In fact Pharo 1.4 is roughly 1 MB smaller than Pharo 1.3.
Even if you consider Zinc as a movement towards "getting bigger", consider an alternative: having library which follows standards, or keep using hacks which were not complete and useful only for most simple cases.
Zinc is an excellent example, because it is fully backward compatible. I don't see that with RPackage, SystemAnnouncements, Ring, Shout (before Alan fixed it), with the proposed RB changes, ...
I am also against growing system unless it is necessarily. In 1.3. i added a new 'non-interactive' mode and non-interactive ui manager. And this was necessarily, because we need a way to deal with "hanging images" in headless mode. Now we can run images on jenkins, knowing that it will never enter 'click ok to proceed' state.
Yes, this is cool.
Do you think the Pharo we have is good enough to have a future?
No, there is a lot to be improved. I think the future of Pharo is what can be built on top, not what can be integrated and forced upon everybody.
Lukas, there is nobody forcing anyone. How we (or anyone else) can force people to use Pharo?
Right, but you can easily force people not to follow. I don't understand why all the people that still use Pharo 1.0 or 1.1 don't speak up?
A new and 'unproven' tools are 'forced' into images mainly to collect feedback. And i think it is good strategy, because most people *including me* are too busy/lazy to go & download and install new stuff and try it out.
Don't we have all these automatic builds with all these tools pre-loaded? There could be a list of links to these builds on the website and like/+1/report buttons. That would likely yield some discussions and give some time to investigate ...
Lukas
-- Lukas Renggli www.lukas-renggli.ch
-- Jorge Ressia www.jorgeressia.com
Look at Scanner reference to get a feel! I do not call that simple, stable at all. Stef
No more for me.
Introducing more code into Pharo that depends on more parts of Pharo (RPackage, Announcement, Pragma, Ring, RB, Shout, ...) doesn't make it easier to maintain and change Pharo. Or did I misunderstand something about cohesion and coupling? :-)
With that philosophy, we can just declare Pharo as finished and do something else. My point of view is that I invent new abstractions *and than use them* on and for the system itself.
I disagree; I would like a small and stable Pharo in which crazy ideas can be realized. For that I don't need fancy abstractions, but a minimal, simple and absolutely stable system in which I can load and do whatever I want. Maybe this is just me?
Is the current system simple and minimal?
Do you think the Pharo we have is good enough to have a future?
Marcus
-- Marcus Denker -- http://marcusdenker.de
participants (5)
-
Igor Stasenko -
Jorge Ressia -
Lukas Renggli -
Marcus Denker -
Stéphane Ducasse