add a setting to make height of the input text box configurable as there are no compelling reasons to make the input box hardcoded to one line or to two lines. I would like to see this as a configurable option and allow the user to decide for him/herself how it likes it. there are a lot of reasons why a user would prefer a one line and another set of reasons why would prefer 2 lines.... Let us not be orwellian on this. I will also open a review board with a proposed patch for this. Alin Reproducible: Always
patch on review board https://git.reviewboard.kde.org/r/110111/
Could you please specify the most convincing reasons for both one and two lines so we can determine if they are convincing enough to warrant a setting? Plus (since I don't currently build KTp from Master): Does your patch offer a switch between one and two lines, or a freely configurable minimum number of lines (like Pidgin does)? If we offer a setting, it should be freely configurable (if there are convincing reasons for either one and two lines, there are probably ones for three or four as well).
the patch offers a freely configurable number of lines. for 1 line I see people who do not want to waste space as the box extends anyhow. people who use small screens. people who like their input line somehow consistent over different apps (konversation offers a one line.) and oses... for 2 lines. accessibility reason... people who do not like tight input boxes. the people who want two lines. as reasons for both are pretty valid and in line with kde philosophy... freedom (that is to the user not to developer iirc), making this configurable is the only correct decision. choosing to offer only one looks to me to be a diktat enforced on users and that without any serious technical issues (took me around 20 minutes to write the patch and I am not known as being a c++/gui programming guru) to justify such a freedom limiting decision. furthermore... to counter argument about the usability expert advice. I was looking around at some application that I used or use and found out there is no definitive answer to one line or two lines... google talk... desktop app is 1 line web app is 2 lines android app is 1 line kopete has stretchable to the pleasure of the user. If google could not decide to unify their design to one or two lines (assuming that they can afford and got the best experts on the issue) I do not see the usability expert argument standing. now to conclude... remember that one year ago was an argument in the community about 1 line is good >1 line is bad... now it seems that the motto is 1 line is bad 2 lines is good... I think all of us know how animal farm finished. Alin
forgot skype in that list... 2 lines
A correction: skype actually trying to pleasure the user with stretchable input box as kopete's. I find this is the best option because it can be easily and quickly adjusted without going in any configuration panels.
I don't agree with the splitter on something that dynamically resizes. If you adjust it when you have content typed.. it's just confusing. Plus it never lines up to a nice even line number so it bounces around. As for config options, this is an option that doesn't matter - people quickly adjust to the resizeable thing. We are trialling 2 lines to see if it's better. It's our job as developers to do the hard work and decide what is best for the user. Adding options is the lazy approach, it was never a technical issue it's about doing the best thing.
Can you (or Thomas) explain the "accessibility reason...." behind having 2 lines? The idea that it's a bigger hit area doesn't really work, because the main chat area acts as a focus proxy. (i.e clicking that selects the input box)
It's not really accessibility, but discoverability. Two lines make it clear that it's a multi-line text field, whereas an initial single line suggests that it is just a single line field, like in the Chat Plasmoid for example. This is not a very strong case, though. People should be fine with one line, two just communicate "Your message length is not limited" more clearly. For me, both variants are okay, none of them should be a problem for any user. Yes, Pidgin introduced a setting, but that doesn't mean it's necessary. It's not about the effort of putting in a setting: It's about taking a more conservative stance on adding new settings in general. We do not want completely unconfigurable applications like Empathy, but we don't want the configuration monstrosities of the KDE3 age either. Some KDE SC applications and some workspace parts are already dangerously close to these times again, but I'm glad KTp is not moving in that direction. That said, probably nobody would die if we introduced that setting, either. The initial height of an auto-expanding box is just not important enough to warrant a setting in my opinion. I agree with David that auto-expand and a splitter do not mix well at all. This is because a splitter suggests "direct manipulation", i.e. if I drag the splitter to position X, I expect it to stay there instead of being pushed up automatically when I type more text. Neither Kopete nor Skype have an auto-expanding box, so for them it works fine. So from my perspective, the options are (in order of preference, 1 is most preferred) 1. Two lines, no setting, auto expand 2. One line, no setting, auto expand 3. Setting for initial number of lines, auto expand 4. Splitter to set height, no auto expand
so you have a case of a feature that has obviously no absolute solution -- you yourself acknowledged that this was already a much divisive and debated issue in another community which was concluded by adding an option and instead of learning from their experience -- and you advise to ignore lessons of history and embrace a bullish attitude... the developer knows better. I gave you above valid use cases for both 1 or 2 lines... and you say that does not warrant an option... In your opinion, can you give some examples of features that do need to have an option? Would be also nice to hear your usability argument for 2 lines on a 11" or smaller screen. Freedom to configure software is not a monstrosity how you put it. I have serious doubts that was one at kde 3.5 time also... So I suspect you are a supporter of this marketing movement that goes around in kde at the moment that choice is bad (that is a dogmatic, anti-liberal and anti-progressive stance from whatever point of view you want to argue). I really fail to see any merit in your argument against configuration. While I may agree with you that presenting all the options in a configure menu as default view is overwhelming, I really do not think that removing the options is the solution -- redesign the way in which they are presented... I will explain you why, for me, is important to have a one line text input. I use konversation which is a one line input and I use gtalk mobile version which is a one line input (on my adroid even the sms input is one line)... changing to a two line input is a real nuisance(gtalk offers that in their web app, I experienced it already and gave up on it)... For me is simple to make a decision on what application to use if does not offer what I need. If I want to use software build under the motto "we know better what is good for you" I would revert to mac osx that came with the hardware. While I agree that a user that uses only ktp-text-ui for IM can adapt, it is not the case for users like me who use multiple applications. to conclude, text-ui, in the behaviour section, has at the moment 3 options do you really think that adding one more, that offers freedom to users, transforms ktp-text-ui in a monstrosity? if adding a fourth create unimaginable strain on users who can engage no more than 2 neurons at any given moment of time, we can put and advanced tab and there add the liberating features.
I'd like to point out one thing, that has not been raised so far - who are our target users and would they be happy with or without the option? Would they be happy with 1 or 2 lines as default? Answers to these questions should drive the final decision. > While I may agree with you that presenting all the options in a configure menu as default view > is overwhelming, I really do not think that removing the options is the solution We are actually not removing the option, there is no option. And just as you said, too many options are overwhelming. Soon we'll be arguing for 5th option to have a send button...6th option to configure the content of context menu, because, why not? Yes, we as developers sometimes do know better. After all, it's our job and we do it, well, because we do know better. I'm with Thomas on this one.
and I install kopete exactly in this moment...
Martin, do you really mean to say you will define some imaginary "target users" thus excluding anyone that does not match? Seriously? How many users does ktp have right now by your estimate? Even if you are generous do you really think this sample is representative of the KDE base at large? You say being a developers magically makes you know better what the user shall see, well let me use your approach and take you for a ride... what next, if you were plasma devs would you force me to use widgets of certain size only?? cause I may resize it to a rational ration rather than strictly integer?? Hell... you shall really read Orwell I for example feel crammed and claustrophobic in the tight auto-resizing box on the desktop. I use kpt on a Nokia phone and it is fine being 1 line there, I am guessing the box frame being very very think makes a difference. This is an issue of taste too and if there is no technical reason as you acknowledge above then what is the problem of having it as extra option in 'advanced' section of sorts? None of the arguments answers this instead you say you recognise the particularity of the situation but will go ahead bullishly anyway 'cause users will get used to it'? As Alin very appropriately appropriately pointed out one of the main attractions of KDE is its configurability. If users are to get used to whatever the devs decide then whats the point of KDE? There are 2 GNOMEs already covering this, and for those not overly concerned with politics there is Mac OS X.
> Martin, do you really mean to say you will define some imaginary "target users" thus excluding anyone that does not match? Seriously? Yes, that is how you do actual software development, by defining the end users and their needs, followed by use-cases. We actually even have them defined, see here: http://community.kde.org/KTp/Usability Also note, that these are very similar to the rest of KDE personas. > what next, if you were plasma devs would you force me to use widgets of certain size only FYI, there will be a grid in the next plasma2 and yes, that will let you resize only to a certain sizes. Sorry for the bad news. > Hell... you shall really read Orwell And some reading for you: http://www.kde.org/code-of-conduct/ > If users are to get used to whatever the devs decide then whats the point of KDE? KDE is not about developers having to provide options for every single bits of UI and every single user out there. This is nowhere written. We decided how the contact list is layouted. People got used to it. We decided what colors will text-ui tabs have when the user is typing. People got used to it. And I can give you many more examples. So yes, we can make a decision and yes people will get used to it. It works. As for this particular case, Thomas is a usability expert, I trust him. Many things in KTp are the result of his advices and it just works(tm).
http://community.kde.org/KTp/Usability : Almost "no comment", i guess someone probably thought it was funny. Cool! It is not my cup of tea and won't judge it. If you however have actual thoughts on how to define these target users I am interested. From the code of conduct: 1. Be considerate 2. Be respectful 3. Be collaborative 4. Be pragmatic 5. Support others in the community 6. Get support from others in the community With all respect it is not the place here to speak about literature so I will not elaborate on my sincere recommendation. Your totalitarian approach however does breach a number of the stated principles in the code of conduct, at least 1, 3, 4, 5, 6. You are giving example with colours, to this I can only say you haven't got any number of users so far to be able to say whether they go on with it or not. I stated my reasons for preferring larger box, Alin stated arguments for both choices, as mentioned already there was a heated exchange in pidgin over this issue so I hope you understand not all people share the same preferences for little things like this. I am OK with the colours but I can assure you if someone really does not like them then discovers it is impossible to change them chances are you are not going to hear from them unless they are already invested. When ktp grows and many people start using it this kind of thing thing will not fly. A quick consultation with google told me one can change the colours in iChat for example. But enough with specifics. Now what is really the problem with inclusion of a minority of people who have different perception of beauty? How is excluding them helping anything? You may say KDE3 style configuration hell, well if this ever existed how about simply not showing all configurations? It looks a lot more considerate, respectful, inclusive, collaborative, pragmatic etc.. to me. You are not going to look at my screen and typing in my input box all day every day neither am I am going to dictate how you use your chat window. If the problem is not technical as you said then how is such covert inclusion not an usability improvement?
> Almost "no comment", i guess someone probably thought it was funny. Again, that's how development works. Feel free to contact KDE's usability team to give you more details about what is user-centered development. Then you will understand what is that about. > I can only say you haven't got any number of users so far to be able to say whether they go on with it or not. On what grounds you base the claim that we "don't have enough users"? I'd be actually interested to see some numbers about our users. > so I hope you understand not all people share the same preferences for little things like this And we will never ever make every single user happy. That is life. > When ktp grows and many people start using it this kind of thing thing will not fly. FYI, we're the default client on many distros. So far nobody complained about missing config for tab colors. > A quick consultation with google told me one can change the colours in iChat for example. Right. Does it allow you to set size of the input field? > neither am I am going to dictate how you use your chat window Yet you dictate how we should do our development. -- I'm still yet to hear, if the super normal/casual user (Penny) will even care about this option. (and with this I'm withdrawing from this discussion as it turned into personal crusades rather than anything else and my opinion is becoming quite biased/clouded)
On the http://community.kde.org/KTp/Usability . If you claim this is an actual guidance then I will comment. It is extreme stereotyping and quite misogynistic that it is offensive indeed. I hope the author does realise there is huge spectrum of personalities. There isn't even a single axis as these 2 points may suggest. It ignores the significant number of intelligent and curious users who like to make their environment comfortable. Let me introduce you to Lena. An intelligent and curious person whose job is a designer and who likes to customise her desktop to her liking. The 'advanced' tab in the KDE settings lets her make her desktop truly hers. Discovering a new program allowing her to call friends on gtalk and facebook from a single place, makes her quite enthusiastic. The enthusiasm however does not last long after she realises the colours and proportions are not customisable. She does not want to deal with people who thought noone would like to change the colours and much less with people who thought that they shall not allow people who want to change the colours. Her rationale being that if the developers followed the code of conduct and a well designed usability setup they would have certainly had these settings available in some intuitive way or at least in the 'advanced' tab. To continue the soap opera: Leonard meets Lena through a common friend and is stunned.... Penny is initially kind of happy not to receive Leonard's spam.... until she sees the fancy coloured iChat of one of her fiends. Martin, I hope you recognise the bias in your opinion and remove it so we can have a rational discussion. I do not dictate how you shall develop but only point out faults with your justification. What do you think the normal user Lena will think of people who think everyone shall wear the same clothes and even go further by trying to justify it with some artificial exaggerated stereotypes? Lets not make things personal but concentrate on the facts. I am glad you realise you are getting more biased and cloudy, and I would very much appreciate if you come back to reason. The arguments are laid out. There is no technical difficulty to put the features. The setting need not be shown to Penny so whether she cares about them of not is irrelevant. She will not be overwhelmed by the settings because she will not see them but when she becomes curious about them she will find them. Lena will likely not raise voice when there are no settings but instead may simply not use the program. Why loose a potential contributor as her over such a silly thing?
PS. the input box in iChat apparently can be set in some obscure configuration.
That's exactly the point of the persona's to work out who we don't cater for. You've correctly analysed that this person would want to configure X and configure Y. This person would be better off with PSI or Vacuum or some such. Trying to cater for all people will result in worse software for everyone. This isn't so much about this particular option, but about setting a precedent of what options we have, and avoid config proliferation. I have not seen any evidence that you understand the arguments that have been laid down against adding the option, and seem to have a more options = more people happy attitude, which is something we know not be true. (http://headrush.typepad.com/creating_passionate_users/2007/02/how_much_contro.html) I am the maintainer of the text-ui and I can override you or Martin at any moment. Frankly I don't think the option as needed, but I probably wouldn't have blocked it either, Alin can testify I've allowed other options that I didn't agree with (log manager scrollback count as an example) after we had a polite discussion. I'm quite laid back :) However since the start the supporters of this config have been mostly focussed on hostility, Orwell references and condescension. The original IRC conversation that predates this was well out of line, and this has been a continuation. I don't like having that attitude in my community, which is why I am closing this.
*** Bug 325407 has been marked as a duplicate of this bug. ***
Hi... just for the record, I'm fine with 1 or 2 lines. But if I should choose, I'd go with one line, auto expand (option 2 in Thomas Pfeiffer comment). However, it's no problem. Hope it gets reverted. Cheers and thanks for KTP!