… which of course, it does not means new tools does not need enhancement/adjustments, and we all can collaborate by reporting issues, and ideas of what we would like of it. but yes, it needs a change of mindset to adopt them. Esteban

On 02 Oct 2014, at 17:34, Alexandre Bergel <alexandre.bergel@me.com> wrote:

Hi Kilon,

Yesterday, at a conference in software engineering, someone shown a new programming environment and made an experiment with users. The authors have shown that users with the _most experience_ in programming technologies are the most negative users. On the opposite, users with less experience in programming environment enjoyed it very much.

Your email clearly goes in the first pattern.

At the beginning I was also in the same situation: I could not make up my mind with GT. Including GT in Pharo is really what could happen the best of our future.

Alexandre

On Oct 2, 2014, at 12:19 PM, kilon alios <kilon.alios@gmail.com> wrote:

...

Is it just me or is the Playground a big step back in the usability of Pharo ?

For me its Gui makes no sense. Gone is the right click menu which had tons of useful actions now you only get like a few options. The tab thing is weird to say the least with the navigation of the tabs being in the opposite side (bottom) of the navigation of the internal tabs(top). No more right arrow menu for shortcuts and many features that workspace had. No ability to resize the panel tabs . Weird terminolgy like "All ref" and "Meta". Meta tab you have to select the first item and then the other first item and then it displays fully .

Also how one closes a tab panel ? Is it possible ? I tried everything but i cant get them to close.

I click a method and instead of opening a panel on the bottom for the source as I would expect ,make my workspace area disappear and create a new panel that sends me to state, whatever this is. Then I have to click on the source tab. This is one of the most annoying guis I have used in an IDE.

Ironically the meta tab has got this right , name of class - method names and source of methods . On the instance side tab there is a different kind of thinking , why ?

Also the inspector seems to get in some kind of loop opening forever panels when it finds a dead end in its inspection.For example clicking self.

It has some good ideas in it, but I dont like the execution.

The big loss for me is the inability to create windows groups, this functionality provided a very nice solution to the windows mess of pharo. I can name my groups but also the individual windows and so the tabs have short names that I know what they are and names of course I can customise.

I can disable it via settings but maybe I am missing something here, a some kind of usefulness because I don't understand or don't know how to use it properly.

-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

Hi Kilon, I think I see where the confusion starts from. So, let's start from the basics: - the inspector is for inspecting objects, not just classes. - every pane shows exactly one object (the one you see in the title of the pane). - every pane shows multiple presentations of that object as tabs. - the two basic presentations are "State" showing the state of the object, and "Meta" showing the code related to the object. We could call this one "Class", but that can generate confusions in certain edge cases when you are inspecting a class object. That is why I chose "Meta", but maybe we can find a better one. - the "Meta" presentation is meant to offer a way to manipulate or lookup the code while being in the context of an object. - every tab is defined by the class of the object you are inspecting. For example, "State" is defined in Object>>#gtInspectorStateIn:, but any other object can define extra tabs. - the "Methods" tab that you refer to is such a presentation that is specific to the class object. Specifically, you can find it in Class>>#gtInspectorMethodsIn:. As the class object is a meta-object, I can see how it can easily generate confusion because you have both "Meta" which will show you the code of the class side of the class, and "Methods" which show you the methods of the class. - that is why, I suggest to start getting used with the inspector by inspecting "FileSystem workingDirectory". You will see that you get and "Items" view that allows you to explore the directory like you do in a more classic browser. Then select from the file system a PNG file and you will get a "Picture" - all these tabs are customizable in that you can define such tabs for your own objects easily. - so, if you do not like one, you can simply either change it or remove it by deleting the corresponding method or removing the annotation from it (gtInspectorPresentationOrder:) Does this make more sense now? As for the actions, they will come back in some form. Cheers, Doru On Thu, Oct 2, 2014 at 10:34 PM, kilon alios <kilon.alios@gmail.com> wrote:

"I will concede that the contextual menu is incomplete. We will work on that. In the meantime, the keybindings work as in the classic workspace.

But, what are the many features that the workspace had and you miss?"

from right click menu I use -> Copy , Paste , find , browse it, sender of it, implementors of it, references to it,. from down arrow menu I use -> Change title, create window group, previous contents (that last on i see that playground has it)

"What do you find weird about "Meta"? I agree that "All ref" should be better explained. But, what do you think these tabs mean?"

Meta means nothing to me, maybe you should name it Meta-Class the problem however is that the IDE does not use that naming anywhere else apart from mentioning it in PBE. No idea what all references means, maybe its the instance variables. Is there an aggreement now in Pharo to call the class side of a class a Metaclass ? Its important to decide on these thing on a global scale or else you have confusion.

"If so, clicking on an entry, is the correspondent of inspecting that object. In the "State" tab, we also provide self to denote the same object as the one represented in the tab. This was borrowed from the classic inspector, but maybe it is confusing in the context of the new inspector. In any case, if you continue clicking on self, you will indeed get the same object over and over. But, that does not mean that the inspector went in a loop"

yes but why allowing the same pane to respawn when clicking self it makes not sense to me. Its also confusing for a begineer.

"But, what do you mean by instance side tab?"

I assume the instance side of the class is the "methods" tab ? no ?

"Which one did you like?"

I like the basic idea of trying to extend workspace with much more functionality and uniting all tools like inspector and system browser under one roof. I like also that comments get a new separate tab and that you trying to support pillar.

The State name is vague, since state can mean both instance and class variables.

"I do not understand this part. Can you explain in more details?"

this -> https://www.youtube.com/watch?v=GGJZeajjWGU

"It is not different thinking. It is a different use case :). The Meta tab shows the class of the object you inspect. If you inspected a class, then the inspector also provides you a list of its methods. "

I am talking about the methods tab which also a confusing name for a tab, what kind of methods ? instance or class . I assume from the existence of Meta that those are instance methods . So why methods tab is not similar to Meta tab ? why it show only the method names and not the class and its inheritance chain and the method source of the method selected from the list of methods exactly as Meta tab is doing ?

-- www.tudorgirba.com "Every thing has its own flow"

What is not clear? For documentation, there are plenty of examples on the humane-assessment.com/blog. Here are some: http://www.humane-assessment.com/blog/the-moldable-gtinspector-deconstructed http://www.humane-assessment.com/blog/browsing-files-with-gtinspector/ http://www.humane-assessment.com/blog/browsing-files-with-gtinspector-video/ http://www.humane-assessment.com/blog/dynamic-exploration-of-a-postgres-db-w... http://www.humane-assessment.com/blog/managing-gtinspector-extensions/ http://www.humane-assessment.com/blog/writing-pillar-books-with-the-gtinspec... http://www.humane-assessment.com/blog/managing-morph-keybindings-with-gtinsp... http://www.humane-assessment.com/blog/introducing-the-gtplayground/ Or you can search https://www.google.com/search?q=site:humane-assessment.com+gtinspector Doru On Fri, Oct 3, 2014 at 8:18 AM, kilon alios <kilon.alios@gmail.com> wrote:

not really but thats ok I will wait for the documentation of it.

On Fri, Oct 3, 2014 at 8:46 AM, Tudor Girba <tudor@tudorgirba.com> wrote:

...

Hi Kilon,

I think I see where the confusion starts from. So, let's start from the basics: - the inspector is for inspecting objects, not just classes. - every pane shows exactly one object (the one you see in the title of the pane). - every pane shows multiple presentations of that object as tabs. - the two basic presentations are "State" showing the state of the object, and "Meta" showing the code related to the object. We could call this one "Class", but that can generate confusions in certain edge cases when you are inspecting a class object. That is why I chose "Meta", but maybe we can find a better one. - the "Meta" presentation is meant to offer a way to manipulate or lookup the code while being in the context of an object. - every tab is defined by the class of the object you are inspecting. For example, "State" is defined in Object>>#gtInspectorStateIn:, but any other object can define extra tabs. - the "Methods" tab that you refer to is such a presentation that is specific to the class object. Specifically, you can find it in Class>>#gtInspectorMethodsIn:. As the class object is a meta-object, I can see how it can easily generate confusion because you have both "Meta" which will show you the code of the class side of the class, and "Methods" which show you the methods of the class. - that is why, I suggest to start getting used with the inspector by inspecting "FileSystem workingDirectory". You will see that you get and "Items" view that allows you to explore the directory like you do in a more classic browser. Then select from the file system a PNG file and you will get a "Picture" - all these tabs are customizable in that you can define such tabs for your own objects easily. - so, if you do not like one, you can simply either change it or remove it by deleting the corresponding method or removing the annotation from it (gtInspectorPresentationOrder:)

Does this make more sense now?

As for the actions, they will come back in some form.

Cheers, Doru

On Thu, Oct 2, 2014 at 10:34 PM, kilon alios <kilon.alios@gmail.com> wrote:

...

"I will concede that the contextual menu is incomplete. We will work on that. In the meantime, the keybindings work as in the classic workspace.

But, what are the many features that the workspace had and you miss?"

from right click menu I use -> Copy , Paste , find , browse it, sender of it, implementors of it, references to it,. from down arrow menu I use -> Change title, create window group, previous contents (that last on i see that playground has it)

"What do you find weird about "Meta"? I agree that "All ref" should be better explained. But, what do you think these tabs mean?"

Meta means nothing to me, maybe you should name it Meta-Class the problem however is that the IDE does not use that naming anywhere else apart from mentioning it in PBE. No idea what all references means, maybe its the instance variables. Is there an aggreement now in Pharo to call the class side of a class a Metaclass ? Its important to decide on these thing on a global scale or else you have confusion.

"If so, clicking on an entry, is the correspondent of inspecting that object. In the "State" tab, we also provide self to denote the same object as the one represented in the tab. This was borrowed from the classic inspector, but maybe it is confusing in the context of the new inspector. In any case, if you continue clicking on self, you will indeed get the same object over and over. But, that does not mean that the inspector went in a loop"

yes but why allowing the same pane to respawn when clicking self it makes not sense to me. Its also confusing for a begineer.

"But, what do you mean by instance side tab?"

I assume the instance side of the class is the "methods" tab ? no ?

"Which one did you like?"

I like the basic idea of trying to extend workspace with much more functionality and uniting all tools like inspector and system browser under one roof. I like also that comments get a new separate tab and that you trying to support pillar.

The State name is vague, since state can mean both instance and class variables.

"I do not understand this part. Can you explain in more details?"

this -> https://www.youtube.com/watch?v=GGJZeajjWGU

"It is not different thinking. It is a different use case :). The Meta tab shows the class of the object you inspect. If you inspected a class, then the inspector also provides you a list of its methods. "

I am talking about the methods tab which also a confusing name for a tab, what kind of methods ? instance or class . I assume from the existence of Meta that those are instance methods . So why methods tab is not similar to Meta tab ? why it show only the method names and not the class and its inheritance chain and the method source of the method selected from the list of methods exactly as Meta tab is doing ?

-- www.tudorgirba.com

"Every thing has its own flow"

-- www.tudorgirba.com "Every thing has its own flow"

It is precisely those points that are not so obvious that are important to discover. That is why your feedback is important. Please keep the questions and suggestions coming. Cheers, Doru On Fri, Oct 3, 2014 at 9:00 AM, kilon alios <kilon.alios@gmail.com> wrote:

Thanks I will read those and try it in practise this way it will be easier to me to connect the dots. I was not aware of them, maybe add them to a chapter to Pharo for the Enterprise to make them more visible , they will also need to be added to PBE for beginners because many things about this tool are not obvious. Still a long way till Pharo 4 is release so you have plenty of time to polish things up.

​

-- www.tudorgirba.com "Every thing has its own flow"

Tudor Girba wrote:

and "Meta" showing the _code_ related to the object. [..] but maybe we can find a better one.

Maybe "code" for instance side methods ?...

"Meta" which will show you the code of the class side of the class, and "Methods" which show you the methods of the class.

...and "Code" for class side methods ? Or "methods" and "Methods" for instance-side and class-side respectively, or "Code" and "MetaCode" ? -ben

On 02 Oct 2014, at 21:36, Tudor Girba <tudor@tudorgirba.com> wrote:

On Thu, Oct 2, 2014 at 5:19 PM, kilon alios <kilon.alios@gmail.com> wrote: Is it just me or is the Playground a big step back in the usability of Pharo ?

For me its Gui makes no sense. Gone is the right click menu which had tons of useful actions now you only get like a few options. The tab thing is weird to say the least with the navigation of the tabs being in the opposite side (bottom) of the navigation of the internal tabs(top). No more right arrow menu for shortcuts and many features that workspace had.

I will concede that the contextual menu is incomplete. We will work on that. In the meantime, the keybindings work as in the classic workspace.

And one needs to take into account that the “pull” strategy for menu items is much better than a “push”: Instead of adding just everything that was there, wait till people get *really* upset and implement it. This proves that they *really* need it :-) In a way we used that strategy *a lot* for Pharo… we removed so many menu entries everywhere. Or Settings… the amount of settings that I removed, I guess >100? And the fun thing is: nobody ever complained :-) But if I would have asked: you can be sure that someone reacts like “I never used it but now that you tell me, I want to keep it”. (this is true for every unused feature. Just remove and wait. Never ask. If you ask you will have to keep it. People react to removing unused features as if you would take something away from them that was just stored securely for bad times, as if it would be deeply ingrained in our genes or something). Marcus

Well, I am COMPLAINING LOUDLY about this lack of menu entries then. I hate this super shrunk list of entries in the playground and the GT toolset. I love GT, but I hate those damn menus that do not have what I do need (which is give me the full menu where I do use the same options Kilon mentioned). Also, on Cmd-o used to open about all tools, yes, Cmd-o Cmd-p --> Monticello Cmd-o Cmd-w --> Workspace Cmd-o Cmd-b --> Browser etc. Just GTInspect KMCategory allInstances and see the two global shortcuts categories (one for Spotlight, one for the global shortcuts - and yeah, they should have a name displayed other than nil). On my own little Pharo, I've set addtional shortcuts for the Finder, the File Browser etc as these are sorely missing. And GT as well. Cmd-o Cmd-g for opening a Play<g>round. So that I can keep Cmd-o Cmd-w for opening a Workspace and keeping all of its abilities intact (and not feel crippled by those replacements). I actually like Esteban's proposal for Cmd-g as what we do with the playground is "give it a go", which is different from a doIt, which can stay local and not have side effects on the right side. Phil On Fri, Oct 3, 2014 at 8:20 AM, Marcus Denker <marcus.denker@inria.fr> wrote:

On 02 Oct 2014, at 21:36, Tudor Girba <tudor@tudorgirba.com> wrote:

On Thu, Oct 2, 2014 at 5:19 PM, kilon alios <kilon.alios@gmail.com> wrote:

...

Is it just me or is the Playground a big step back in the usability of Pharo ?

For me its Gui makes no sense. Gone is the right click menu which had tons of useful actions now you only get like a few options. The tab thing is weird to say the least with the navigation of the tabs being in the opposite side (bottom) of the navigation of the internal tabs(top). No more right arrow menu for shortcuts and many features that workspace had.

I will concede that the contextual menu is incomplete. We will work on that. In the meantime, the keybindings work as in the classic workspace.

And one needs to take into account that the “pull” strategy for menu items is much better than a “push”: Instead of adding just everything that was there, wait till people get *really* upset and implement it. This proves that they *really* need it :-)

In a way we used that strategy *a lot* for Pharo… we removed so many menu entries everywhere. Or Settings… the amount of settings that I removed, I guess >100?

And the fun thing is: nobody ever complained :-)

But if I would have asked: you can be sure that someone reacts like “I never used it but now that you tell me, I want to keep it”. (this is true for every unused feature. Just remove and wait. Never ask. If you ask you will have to keep it. People react to removing unused features as if you would take something away from them that was just stored securely for bad times, as if it would be deeply ingrained in our genes or something).

Marcus

Frankly, I see fogbugz issues closed with some ignore/cannot reproduce status, so, I am not creating any of them anymore. Why bother, except for blocking bugs? I have a set of adjustments on my side and work with them. https://gist.github.com/philippeback is where I store them, an in HOExtras in my SmalltalkHub account. Sad attitude, but I tell you how it is for me. Phil On Fri, Oct 3, 2014 at 12:10 PM, Esteban Lorenzano <estebanlm@gmail.com> wrote:

and where is your enhance proposal? I made a fogbugz search and I didn’t find anything related to improve “cmd+o"

Esteban

On 03 Oct 2014, at 09:28, phil@highoctane.be wrote:

Well, I am COMPLAINING LOUDLY about this lack of menu entries then.

I hate this super shrunk list of entries in the playground and the GT toolset. I love GT, but I hate those damn menus that do not have what I do need (which is give me the full menu where I do use the same options Kilon mentioned).

Also, on Cmd-o used to open about all tools, yes, Cmd-o Cmd-p --> Monticello

Cmd-o Cmd-w --> Workspace

Cmd-o Cmd-b --> Browser

etc.

Just GTInspect

KMCategory allInstances

and see the two global shortcuts categories (one for Spotlight, one for the global shortcuts - and yeah, they should have a name displayed other than nil).

On my own little Pharo, I've set addtional shortcuts for the Finder, the File Browser etc as these are sorely missing.

And GT as well. Cmd-o Cmd-g for opening a Play<g>round. So that I can keep Cmd-o Cmd-w for opening a Workspace and keeping all of its abilities intact (and not feel crippled by those replacements).

I actually like Esteban's proposal for Cmd-g as what we do with the playground is "give it a go", which is different from a doIt, which can stay local and not have side effects on the right side.

Phil

On Fri, Oct 3, 2014 at 8:20 AM, Marcus Denker <marcus.denker@inria.fr> wrote:

...

On 02 Oct 2014, at 21:36, Tudor Girba <tudor@tudorgirba.com> wrote:

On Thu, Oct 2, 2014 at 5:19 PM, kilon alios <kilon.alios@gmail.com> wrote:

...

Is it just me or is the Playground a big step back in the usability of Pharo ?

For me its Gui makes no sense. Gone is the right click menu which had tons of useful actions now you only get like a few options. The tab thing is weird to say the least with the navigation of the tabs being in the opposite side (bottom) of the navigation of the internal tabs(top). No more right arrow menu for shortcuts and many features that workspace had.

I will concede that the contextual menu is incomplete. We will work on that. In the meantime, the keybindings work as in the classic workspace.

And one needs to take into account that the “pull” strategy for menu items is much better than a “push”: Instead of adding just everything that was there, wait till people get *really* upset and implement it. This proves that they *really* need it :-)

In a way we used that strategy *a lot* for Pharo… we removed so many menu entries everywhere. Or Settings… the amount of settings that I removed, I guess >100?

And the fun thing is: nobody ever complained :-)

But if I would have asked: you can be sure that someone reacts like “I never used it but now that you tell me, I want to keep it”. (this is true for every unused feature. Just remove and wait. Never ask. If you ask you will have to keep it. People react to removing unused features as if you would take something away from them that was just stored securely for bad times, as if it would be deeply ingrained in our genes or something).

Marcus

On 03 Oct 2014, at 12:21, phil@highoctane.be wrote:

Frankly, I see fogbugz issues closed with some ignore/cannot reproduce status, so, I am not creating any of them anymore. Why bother, except for blocking bugs?

Tell me: What is the difference between a bug that it fixed and one that can not be reproduced?

I have a set of adjustments on my side and work with them.

https://gist.github.com/philippeback is where I store them, an in HOExtras in my SmalltalkHub account.

Sad attitude, but I tell you how it is for me.

I think I should take the same attitude. I will just fix stuff for me from now on. Marcus

One that cannot be reproduced right off the bat is one which is more complex than meets the eye. Heisenbugs? Still, bugs. Easy to say, of course. Now, maybe putting things as "closed" when you cannot reproduce is not okay with me. I understand the scale of the issues list and that you do your best. And, as much I do want to help the community, I first want to help me first in delivering projects for clients. That is what will help in putting money back in the community after all. I can live with the current state of things, no issue (pun intended). Phil On Fri, Oct 3, 2014 at 12:52 PM, Marcus Denker <marcus.denker@inria.fr> wrote:

On 03 Oct 2014, at 12:46, Marcus Denker <marcus.denker@inria.fr> wrote:

On 03 Oct 2014, at 12:21, phil@highoctane.be wrote:

Frankly, I see fogbugz issues closed with some ignore/cannot reproduce status, so, I am not creating any of them anymore. Why bother, except for blocking bugs?

Tell me: What is the difference between a bug that it fixed and one that can not be reproduced?

I wonder if people have an idea of the scale of things…

- We closed 107 issues the last 7 days - there are 618 open issues - There where 11268 issues reported for Pharo since we started.

So when I look at an issue i try to reproduce it. If I can not, I close it. Else this is not possible to do, sorry. Even if I wanted to: If we keep all non-reproducible cases open we soon will be paralysed. (And that by issues that to 99% are fixed).

Marcus

I just tell you what the feeling is. This has happened for a lot of what I reported. Do not mistake me, Pharo is usable and I haven't been encountering blocking issues in 3.0. Phil On Fri, Oct 3, 2014 at 12:46 PM, Marcus Denker <marcus.denker@inria.fr> wrote:

On 03 Oct 2014, at 12:21, phil@highoctane.be wrote:

Frankly, I see fogbugz issues closed with some ignore/cannot reproduce status, so, I am not creating any of them anymore. Why bother, except for blocking bugs?

Tell me: What is the difference between a bug that it fixed and one that can not be reproduced?

I have a set of adjustments on my side and work with them.

https://gist.github.com/philippeback is where I store them, an in HOExtras in my SmalltalkHub account.

Sad attitude, but I tell you how it is for me.

I think I should take the same attitude. I will just fix stuff for me from now on.

Marcus

Solution is more engineers. You are not targeted as a person, nor do I expect you to solve this in any way. And you are astounding in how much work you tackle. It is a systemic issue, this is just a symptom. Phil On Fri, Oct 3, 2014 at 1:12 PM, Marcus Denker <marcus.denker@inria.fr> wrote:

On 03 Oct 2014, at 13:03, phil@highoctane.be wrote:

...

I just tell you what the feeling is. This has happened for a lot of what I reported.

And what is the solution? I fixed *HUNDRETS* of issues.

maybe I should stop.

...

Do not mistake me, Pharo is usable and I haven't been encountering blocking issues in 3.0.

yes, maybe this is because some idiot (me) is going through *THOUSANDS* of reports of problems, tries to recreate them, fixes those he can… and close those where I can do nothing so I can focus on the others.

Alternatively, I can just let all the issue tracker entries open. Will that lead to a system that is as stable?

Marcus

Marcus Denker wrote:

On 02 Oct 2014, at 21:36, Tudor Girba <tudor@tudorgirba.com <mailto:tudor@tudorgirba.com>> wrote:

...

On Thu, Oct 2, 2014 at 5:19 PM, kilon alios <kilon.alios@gmail.com <mailto:kilon.alios@gmail.com>> wrote:

Is it just me or is the Playground a big step back in the usability of Pharo ?

For me its Gui makes no sense. Gone is the right click menu which had tons of useful actions now you only get like a few options. The tab thing is weird to say the least with the navigation of the tabs being in the opposite side (bottom) of the navigation of the internal tabs(top). No more right arrow menu for shortcuts and many features that workspace had.

I will concede that the contextual menu is incomplete. We will work on that. In the meantime, the keybindings work as in the classic workspace.

And one needs to take into account that the “pull” strategy for menu items is much better than a “push”: Instead of adding just everything that was there, wait till people get *really* upset and implement it. This proves that they *really* need it :-)

In a way we used that strategy *a lot* for Pharo… we removed so many menu entries everywhere. Or Settings… the amount of settings that I removed, I guess >100?

And the fun thing is: nobody ever complained :-)

But if I would have asked: you can be sure that someone reacts like “I never used it but now that you tell me, I want to keep it”. (this is true for every unused feature. Just remove and wait. Never ask. If you ask you will have to keep it. People react to removing unused features as if you would take something away from them that was just stored securely for bad times, as if it would be deeply ingrained in our genes or something).

Marcus

I think that is a fair strategy, as long as attention is paid when the complaints come. In my day job, the project managers don't like us just asking our (internal) clients "do you want this" - since they _always_ just say Yes without having to consider project budget. The project then grows until it is too expensive to pass a cost/benefit analysis at the executive toll-gate. -ben