On 9 January 2014 14:29, Esteban A. Maringolo <[hidden email]> wrote: From time to time, a slippery road warning sign is good even for the I like your analogy with roads and driver. Let me put mine, which i think fits better for modal dialogs: https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcR1S9qdN4wl00uaCdNa70_IxhfyodgQGzwkNCMIADlgJmxFcXFgPA Esteban A. Maringolo 2014/1/9 Igor Stasenko <[hidden email]>: > > > > On 9 January 2014 14:03, Marcus Denker <[hidden email]> wrote: >> >> >> On 09 Jan 2014, at 13:59, kilon alios <[hidden email]> wrote: >> >> > yes this lead to endless debate, but I think this is good because we see >> > diffirent ways into looking into things. I have to admit till today I never >> > expected that someone would be against confirm dialogs to such extend of >> > wanting them to be removed completely, but I can see now that for people >> > that dont make mistakes or they rather live with these mistakes would prefer >> > a non confirmation approach. Its good to discuss these things because next >> > time I will try to "fix" something I will try to do it in a way that pleases >> > most people and not interrupting their workflow. >> > >> > And its endless because people prefer diffirent things, and thats ok. >> > Opinions should be expressed and be respected. Opinions matter to make >> > software better . Afterall software is made to please people by doing the >> > things they want . No software of course is perfect. :) >> > >> > I completely respect Igor's opinion. And Igor its great you have worked >> > on these things and thank you :) You should promote your work I think >> > progress should be more carefully logged so we can all appreciate the work >> > that goes inside pharo :) >> > >> >> A editor key combination that looses code because is does not support undo >> is *wrong*. Completely and utterly *wrong*. >> > > Crippling the feature is *wrong*. Completely and utterly *wrong*. > As we all agreed , the proper fix to cmd-l problem is make it undoable. > An *improper* fix is put warnings everywhere. Warnings do not prevent from > mistakes, > they just do things worse at times. > That's all what wanted to say. > >> >> >> Marcus > > > > > -- > Best regards, > Igor Stasenko. -- Best regards, Igor Stasenko. |
the solution is to set that on settings
use barriers for protection and let hackers do its risky thing Pharo will be better if Igor can use it like a ninja and the rest of us noobs not so eager to loose code changes stick to a safer workflow On Jan 9, 2014, at 1:50 PM, Igor Stasenko <[hidden email]> wrote:
|
In reply to this post by Igor Stasenko
2014/1/9 Igor Stasenko <[hidden email]>:
> On 9 January 2014 14:29, Esteban A. Maringolo <[hidden email]> wrote: >> >> From time to time, a slippery road warning sign is good even for the >> best driver. >> > I like your analogy with roads and driver. Let me put mine, which i think > fits better for modal dialogs: > https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcR1S9qdN4wl00uaCdNa70_IxhfyodgQGzwkNCMIADlgJmxFcXFgPA :) In a normal world road signs are placed beside or above the road. Never in the middle, unless there is something exceptional ahead, like an accident that can cause you to lose your live, or your code (which sometimes feels alike :P) Esteban A. Maringolo |
In reply to this post by philippeback
Hi all -- I didn't read this discussion closely, I just saw this screenshot and had to say something. I do not wish to criticize Varicad since it's evident the developers put a lot of work into the domain functionality.. .. But this UI design provides way too-low "functional-density". They're letting "commands" drive the UI-design rather than domain objects. Those icon images are probably burning into their monitor screens and eating up valuable screen space that could otherwise be used to present the domain in a less distracting fashion (i.e., with contextually available commands appearing on-demand).
I'm glad to see Pharo regarding this as an example of what _not_ to do. :) On Thu, Jan 9, 2014 at 5:11 AM, [hidden email] <[hidden email]> wrote:
|
About Varicad: Check the bottom left area saying "command". Any icon is doing some command thing. So, once you know them, you type them in the box. Kind of DSL thing.
For $650, there is a lot in the package. Competition are in the thousands. Google Sketchup is not up to anything, nor is open source in that area. Phil On Thu, Jan 9, 2014 at 6:05 PM, Chris Muller <[hidden email]> wrote:
|
In reply to this post by Igor Stasenko
Igor Stasenko wrote:
> and guess what i currently working on? > OSWindow interface, which will allow us to control creating of windows > and managing them from the image, and handling the events. > OSWindow sounds really exciting. Perhaps it will help facilitate multi-monitor support? cheers -ben |
On 11 January 2014 15:20, <[hidden email]> wrote:
eventually, yes. at least we care about it from very beginning: Object subclass: #OSWindowAttributes instanceVariableNames: 'bounds title fullscreen screenId preferableDriver' classVariableNames: 'DefaultBounds DefaultFullscreen DefaultTitle' poolDictionaries: '' category: 'OSWindow' this object is used to describe a window attributes. sure thing not all of them listed there and much more will be added later, but as you can see, screenId is already there, so potentially you can control which screen your window must use. it is not clear now, how we will facilitate the screen selection and what values/api to use to identify them, but at this early stage we lying a foundations to make this option come into play one day. cheers -ben -- Best regards, Igor Stasenko. |
I finally read this whole thread. I don't normally get involved in
Pharo discussions but, I can't help to say: Igor is 100% right, and his best rationale why at the beginning: ========== at the end, it is just silly: the point is that i am always sure about things i do, when interacting with my computer. if i'm not, i simply don't (or i save and then proceed with caution).. and in any case, i don't need stupid UI asking me about things i already decided to do (yes i am sort of man, who disables file deletion warning, if it provided). =========== Unless there's a possibility of an accidental mishap with great consequences, software should simply do what its told without question. kilos doesn't want to be punished for a "mistake". Much simpler, if you don't want to "lose your changes", is to simply not press Command+L in the first place. It sounds like kilos is pressing buttons before knowing the effects. That's just impatience that should be appropriately spanked as Igor said. What if, upon being asked, "are you sure" you made your mistake right then and there, by pressing "Yes" instead of "No"? This is very easy to happen since, as Igor pointed out, mechanical reactions to pop-ups develop over time.. I disagreed with, "A editor key combination that looses code because is does not support undo is *wrong*." Losing the code is the _entire purpose_ of Command+L. Every version of every method is still saved in history so one could simply save their work then revert to the prior version to avoid "losing code". Undo on Command+L would be of no use except for people pressing buttons before thinking. That needs to be corrected to, "think first, THEN press buttons." :) Bye. On Sat, Jan 11, 2014 at 8:33 AM, Igor Stasenko <[hidden email]> wrote: > > > > On 11 January 2014 15:20, <[hidden email]> wrote: >> >> Igor Stasenko wrote: >>> >>> and guess what i currently working on? >>> OSWindow interface, which will allow us to control creating of windows >>> and managing them from the image, and handling the events. >>> >> OSWindow sounds really exciting. Perhaps it will help facilitate >> multi-monitor support? >> > eventually, yes. > at least we care about it from very beginning: > > Object subclass: #OSWindowAttributes > instanceVariableNames: 'bounds title fullscreen screenId > preferableDriver' > classVariableNames: 'DefaultBounds DefaultFullscreen DefaultTitle' > poolDictionaries: '' > category: 'OSWindow' > > this object is used to describe a window attributes. sure thing not all of > them > listed there and much more will be added later, but as you can see, > screenId is already there, so potentially you can control which screen your > window must use. > it is not clear now, how we will facilitate the screen selection and what > values/api to use to identify them, but at this early stage we lying a > foundations to make this option come into play one day. > > >> >> cheers -ben >> > > > > -- > Best regards, > Igor Stasenko. |
i have an impression, that people simply don't understand the function of cmd-l, in debugger i more often operate with method's source pane than inspectors:and how convenient it is, especially when debugging: inspecting result, simply i pressing Cmd-L and ready to go... Less often i use Cmd-I and inspector panes, because it takes much more time and more interactions to get what i want. Note, that in such way of use (cmd-p/cmd-l) there is no chance why i would want to save changes instead of thrashing them, and asking me about it is just silly. On 12 January 2014 01:08, Chris Muller <[hidden email]> wrote: I finally read this whole thread. I don't normally get involved in -- Best regards, Igor Stasenko. |
Both saving changes and cancelling changes should be easy to do, without
confirmation. They are both too common an action to slow down in this way. However, *both* actions should be undo-able. Regards, -Martin |
In reply to this post by Chris Muller-3
On Jan 11, 2014, at 10:08 PM, Chris Muller <[hidden email]> wrote:
the point is that i am always sure about That means that you're exceptional Chris! Like in the exception (and not the rule) Undo on Command+L would be of no use except for people pressing Which goes in front collision course with the usability principle "Don't make me think" for designing great user interfaces If smalltalk is an environment where experimenting is extremely cheap and feedback is instant, there is not reason to overload the neocortex of the user forcing her to always be sure (in anticipation of her key strokes!). If you do it, it will be one step in the direction of making an unexperienced user feel unsafe or worst excluded from the smalltalk ecosystem (because other ecosystems do think in making newcomers feel safe). BTW, experiments where you aren't completely sure of the outcome are often surprising, fun and source of lots of insights, aren't they? |
On 12 January 2014 15:52, Sebastian Sastre <[hidden email]> wrote:
add me to that exception. you got 2 exceptions now.
oh come on... you know, if you so care about overloading neocortex,
then i think you should choose different occupation in your life:
there's many which don't requires too much thinking, in contrary to
programming. to feel safe you need to know what you doing, and predict consequences.. and believe me or not, this involves some thinking.
pressing random keys without any clue what you doing is exciting activity, i guess, but it is far from what people calling "programming" or "coding". Cheers :) -- Best regards, Igor Stasenko. |
In reply to this post by sebastianconcept@gmail.co
On 12 January 2014 15:52, Sebastian Sastre <[hidden email]> wrote: How do you see it? Like:
Pharo - programming language that don't makes you think? Sorry, without me then. -- Best regards, Igor Stasenko. |
On 12 Jan 2014, at 13:54, Igor Stasenko <[hidden email]> wrote:
> How do you see it? Like: > Pharo - programming language that don't makes you think? You are mixing the language and the IDE here Ben |
On 12 January 2014 17:57, Benjamin <[hidden email]> wrote:
I'm not mixing anything.. IDE is part of it. more than anything else. But most funniest part of it, that yes/no popups going against such principle, because when you see popup, it forces you to a) read what it says b) understand the question c) make decision and answer it tell me which part of it "does not makes you think"
:) Ben -- Best regards, Igor Stasenko. |
Administrator
|
In reply to this post by Igor Stasenko
Ha ha ha... this made the whole thread for me! Seriously, I think we all want the same two things: - the system staying out of our way - an opportunity to go back if we make a wrong decision The disagreement seems only to be which item to prioritize *now*, while we're waiting for a solution with both to be developed. And both opinions are reasonable when looked at from that workflow/personality, and seem equally crazy when viewed from the other side. The quickest compromise seems to be making the pop-ups a setting. Here's an issue: Case 12629: Make "discard edits" Confirmation Popups a Setting https://pharo.fogbugz.com/default.asp?12629 Can we all get back to work now? ;)
Cheers,
Sean |
On 12 January 2014 18:49, Sean P. DeNigris <[hidden email]> wrote: Igor Stasenko wrote Sean, please understand, i ranting not because there's something not works now or working not like we would like it to work. I am ranting against approach of using modal popups as way to improve or fill gaps in UI. I also ranting against delusion that "clever" software in some magical way can help preventing mistakes. In my opinion this is achievable only if it will prevent us from doing anything.
-- Best regards, Igor Stasenko. |
Administrator
|
I understand and agree. And that position is not revolutionary - it's a clear case of poor UI design. And, in this case, I think it's easy enough to lose code ( albeit a maximum of 7 lines, right ;-) ) by accidentally closing a browser. Say cleaning the world of multiple unneeded tools and repeatedly typing cmd+w, forgetting that there is unsaved code in 1. Or (which has happened to me) trying to close a bunch of error windows that appeared, and accidentally closing another window because the window order is not what I expected chronologically. So, only in the absence of undo for what IMHO is a likely error, especially for new users, am I suggesting to keep the pop-ups (which I personally hate and would rather lose code occasionally then be interrupted all the time) available for those who want them.
Cheers,
Sean |
Sorry, but it feels there are two discussion mixed:
- if there is an unaccepted edit, it should not be lost, so we need a dialog, undo or history - if you explicitly cancel an edit, there should not be a dialog, because you just said you want to cancel Logic. That is how I understand what Igor says, and that is also my opinion. On 12 Jan 2014, at 19:43, Sean P. DeNigris <[hidden email]> wrote: > Igor Stasenko wrote >> I am ranting against approach of using modal popups as way to improve or >> fill gaps in UI. > > I understand and agree. And that position is not revolutionary - it's a > clear case of poor UI design. > > And, in this case, I think it's easy enough to lose code ( albeit a maximum > of 7 lines, right ;-) ) by accidentally closing a browser. Say cleaning the > world of multiple unneeded tools and repeatedly typing cmd+w, forgetting > that there is unsaved code in 1. Or (which has happened to me) trying to > close a bunch of error windows that appeared, and accidentally closing > another window because the window order is not what I expected > chronologically. > > So, only in the absence of undo for what IMHO is a likely error, especially > for new users, am I suggesting to keep the pop-ups (which I personally hate > and would rather lose code occasionally then be interrupted all the time) > available for those who want them. > > > > ----- > Cheers, > Sean > -- > View this message in context: http://forum.world.st/Do-you-want-to-accept-Discard-tp4735220p4736163.html > Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com. > |
In reply to this post by Igor Stasenko
On Jan 12, 2014, at 2:54 PM, Igor Stasenko <[hidden email]> wrote:
Did I mention usability principle and user interfaces? That's pushing the argument Igor. A) is the UI that makes you think (or not) B) that's good because you have more energy to think in what you actually are trying to achieve |
Free forum by Nabble | Edit this page |