Main » Discussion » GNOME: "Please don't theme our apps" » New reply
    Alert
    You are about to bump an old thread. This is usually a very bad idea. Please think about what you are about to do before you press the Post button.
    New reply
    Post help

    Presentation

    [b]…[/b] — bold type
    [i]…[/i] — italic
    [u]…[/u] — underlined
    [s]…[/s] — strikethrough
    [code]…[/code] — code block
    [spoiler]…[/spoiler] — spoiler block
    [spoiler=…]…[/spoiler]
    [source]…[/source] — colorcoded block, assuming C#
    [source=…]…[/source] — colorcoded block, specific language[which?]
    [abbr=…]…[/abbr] — abbreviation
    [color=…]…[/color] — set text color
    [jest]…[/jest] — you're kidding
    [sarcasm]…[/sarcasm] — you're not kidding

    Links

    [img]http://…[/img] — insert image
    [url]http://…[/url]
    [url=http://…]…[/url]
    >>… — link to post by ID
    [user=##] — link to user's profile by ID

    Quotations

    [quote]…[/quote] — untitled quote
    [quote=…]…[/quote] — "Posted by …"
    [quote="…" id="…"]…[/quote] — ""Post by …" with link by post ID

    Embeds

    [youtube]…[/youtube] — video ID only please
    Thread review
    CaptainJistuce Surely there's an option besides GTK. Qt or something?
    tomman The usual excuses to switching to GTK3 are:

    1) No longer maintained upstream / deprecated
    2) Wayland (which GTK2 does not support)

    1) can be solved with a fork (which noone seems willing to do, unlike what was done with MATE). As for 2), well, I have no interest in Wayland but I can somehow understand that it is "the future". Still, how difficult would be to add Wayland support to GTK2?
    CaptainJistuce Man, I don't get why anyone USES GTK 3.X.

    I know there's the "upgrade from GTK 2 because newer versions are better" angle, but it isn't actually a newer version of GTK 2. It is a fundamentally different thing managed by fundamentally different people, and has proven wildly unreliable for anyone not named Gnome.
    As near as I can tell, no one actually LIKES GTK 3, and it is used solely because people liked GTK 2.
    nyanpasu64 back on topic...
    A few months back, I noticed that all Java Swing apps were missing checkboxes and scrollbars (invisible but clickable). The issue was tracked internally as https://bugs.java.com/bugdatabase/view_bug.do?bug_id=JDK-8218469 .
    From gtk 3.20, gtk has changed the way themes and styles work for many widgets. As per https://developer.gnome.org/gtk3/stable/ch32s10.html, gtk no longer depends on style classes and type names for style matching, but uses element names. Due to which, jdk is not able to render some widgets properly including JSlider. I have tested this by running the SwingSet2 and attached reproducer test.
    Clicking on the link, I read the following:
    The way theming works in GTK+ has been reworked fundamentally, to implement many more CSS features and make themes more expressive. As a result, custom CSS that is shipped with applications and third-party themes will need adjustments.
    Not only does GTK 3.20 break themes, but it breaks applications, and it breaks Java Swing's GTK frontend.
    CaptainJistuce Pretty much. They also harassed Digital Research about Gem, even over petty things like colored title bars.

    And like I said, it takes a certain temerity to sue people for copying your copy.
    Kawaoneechan I'm given to understand when he saw Windows, he said "okay, okay, fair point Bill, but no overlapping windows."
    BearOso
    Posted by CaptainJistuce
    Well, yes, but that's only if you have the financial might of Xerox funding your research, or the temerity to copy Xerox's software and then sue others for ripping you off.

    Windows ripped off Xerox, too, but did a terrible job with it. The original Macintosh team was a close-knit group of truly smart people who cared what they were working on. That's the reason for the better adaptation. That or if Jobs had seen something like Windows, he'd scream "That's shit! Start over!"
    Kawaoneechan Related to the above: https://www.folklore.org/StoryView.py?project=Macintosh&story=A_Rich_Neighbor_Named_Xerox.txt
    CaptainJistuce
    Posted by BearOso
    Posted by creaothceann

    Wasn't the Windows one because of hardware limitations?

    I don’t think so. You can do a lot better with just black and white.
    Well, yes, but that's only if you have the financial might of Xerox funding your research, or the temerity to copy Xerox's software and then sue others for ripping you off.
    BearOso
    Posted by creaothceann

    Wasn't the Windows one because of hardware limitations?

    I don’t think so. You can do a lot better with just black and white.
    creaothceann
    Posted by Screwtape
    I'm pretty sure it's also impossible to have a theme so ugly that everybody hates it; Amiga Workbench 1.0 and Windows 1.0 didn't immediately doom their respective product lines.

    Wasn't the Windows one because of hardware limitations?
    ‮strfry("emanresu")
    Posted by Screwtape
    The thing is, humans hate change, and they are also very, very good at inventing reasonable-sounding justifications for however they happen to feel at any given moment. When some kind of change is announced, it's a safe bet that 90% of the response will boil down to "I hate change" with creative variations. As such, the actual comments themselves are unnecessary; the only interesting information is whether the reactions are significantly more or less than the expected 90% of the total volume.

    Well, not exactly. They'll be negative, but they'll make up some reason for it. And this reason can be more or less bad.

    If the criticism is over really inane stuff ("I don't like the font"), then one should just treat it like praise, kind of like "damned by faint praise" but the other way around. Praised by faint condemnation?

    Supposedly, this is customary in some countries (I think it was China I heard it about). As in, they'll always give negative and only negative feedback, but if it's over some inane stuff that just goes to show there aren't any actual flaws they could find.

    The signals for "this response is not knee-jerk reaction" are things like: longer than average (but not too long), tidy paragraphs instead of sprawling run-on sentences, maybe some formatting like lists and headings, basically anything that indicates the author spent time thinking about the reader instead of scrawling the hottest take they could. None of those things require close reading or empathetic consideration. It's possible some useful and interesting responses get accidentally scooped up with the unhelpful ones, but nothing is ever perfect.

    I'm not sure about this. I've seen perfectly reasonable technical points made in broken English laden with racial slurs, and abject nonsense argued for with very properly written text.

    For the former, just visit any technology forum with sparse moderation. For the latter, just take a look at any Medium post arguing for a stupid idea.
    As for deleting or at least hiding such comments, if someone is serious about listening to interesting feedback, they'll want to come back and re-read that feedback, and show it to other people. That's a lot easier without the unhelpful distractions, even if only because they don't have to scroll so far.

    But the "distractions" they removed didn't seem to have much difference to the critical comments they kept. They were a bit angrier, but they seemed to have about as much merit.

    It's literally impossible to have a theme so perfect that everybody likes it; the best you can hope for is a theme good enough that changing it is more trouble than it's worth. I'm pretty sure it's also impossible to have a theme so ugly that everybody hates it; Amiga Workbench 1.0 and Windows 1.0 didn't immediately doom their respective product lines.

    They don't have to like it, they just have to find it okay. I find those two okay, especially when compared to DOS. A theme designed by God, everyone would find perfect. My point isn't to bring theology into a technological discussion, it's just a thought experiment: if you had a perfect theme, theming support wouldn't be needed.

    And the worse the default theme, the greater the need for theming, and vice versa. Or with your words, the better the theme, the less the ability to change it is worth. So say someone made a truly atrocious default theme: text in #EEFFEE, background #FFEEFF. A theming support for such a distro would be imperative. While if it were the MacOS theme, people would presumably think, "eh, I'd rather have it than not, but this is good enough". So a criticism of the GNOME default theme is definitely relevant and shouldn't be deleted, because the goodness of the default theme has very much to do with the value of theming support.

    A widget toolkit with a theme engine makes some things awesome and some things terrible; a widget toolkit that can't easily be themed (like macOS and Windows have) keeps everything mediocre. Is that better? Maybe? Sometimes?

    Windows does have themes, although you need to modify some system files to use them.
    /nitpick

    More likely, Xfce's default theme is close enough to boring grey-and-blue Adwaita that it doesn't cause any problems. Some Linux vendors have much bolder branding, say in brown and orange, and just to stand out they do unusual things like have menus and toolbars in light-on-dark while the rest of the application is dark-on-light.

    I'm using some theme I downloaded off the internet, "Paper" it's called. One theme in the list ("Orangine") has dark-on-orange menu bars and otherwise dark-on-light. Picking random themes from the list (e.g. light-on-dark themes) doesn't immediately seem to cause any issues. Higan runs on several different platforms with several different themes, I've never heard of any theme breaking its UI.

    GTK+3's theme engine I guess doesn't make it easy to have wildly different colour-schemes for different parts of the window, so themes that want to do that have to add a bunch of custom rules to tweak each supported application... but there's no way to limit a particular rule to a particular application. So if a standard GNOME app happens to name one of its toolbar widgets "switcher", the Linux vendor theme says "widgets named 'switcher' use light-coloured text". Then a third-party app happens to also use the name "switcher" for some other widget that's *not* in the toolbar, so it winds up with light text on a light background.

    Doesn't this imply there's something horribly broken with it if it's so inflexible? They couldn't use something else than CSS for this?

    You do realise that application authors and GTK's maintainers are different people, right?

    But application authors could refuse to ship applications with theming support. For instance, if they'd use imgui, theming wouldn't even be possible.

    To go from "I do not need this thing" to "None of the six billion humans on the planet needs this thing" is quite a feat of extrapolation. Does it genuinely not occur to you that if you don't understand a thing, maybe you just haven't yet encountered the problem it solves?

    Of the seven billion humans on Earth, how many have computers? Of those, how many speak English?

    The "problem" it "solves" has been already been solved by far simpler means such as transcription, English education, and HTML entities. In particular, RTL is completely incomprehensible why it has to be supported on so many levels. Wouldn't it be enough to have the text input convert RTL -> LTR, and then have all the intermediate presentation layers work as usual?

    Case in point, I'm an ESL speaker and I bloody hate localization. Arabs don't even bother writing with Arabic letters, they use the ad-hoc transcriptions of Arabic (e.g. أسعد -> as3ad) because RTL is so god damn wonky (e.g. try pasting "بِسْمِ اللهِ الرَّحْمٰنِ الرَّحِيْمِ" into the terminal and see what happens). Chinese and Japanese often use their local character encodings due to that whole Han unification thing. And for technical discussion almost no matter the language, this is either done in English straight-up (e.g. Linux kernel) or in the local language with almost every single noun replaced by a direct loan from English ("he connected to the server, then liked a post and added the guy who made it to his friends list").

    When you've reached that level, you might as well pull a Torvalds and switch over to English. Then suddenly everything works again. You can now use a spell-checker while writing your documentation, for instance.
    Not strictly relevant to the point under discussion, but I do want to point out PHP's famous "expected T_PAAMAYIM_NEKUDOTAYIM" error.

    Ah yes, the famous "it adds some character to PHP" token.
    I stand by my claim however. T_PAAMAYIM_NEKUDOTAYIM is easy to google, while :: and 4996 are not. Of course T_DOUBLE_COLON would be preferable, but you can't have it all here in life.

    GNU Mailman and INN say hi.

    No mailing list ever embedded me a jQuery.
    Screwtape
    Posted by sureanem
    Well, then you listen to them and see if any of them have reasonable opinions. I mean, you'd have to read through them to censor them anyway, and after you've done that there isn't really much of a point to deleting them.

    The thing is, humans hate change, and they are also very, very good at inventing reasonable-sounding justifications for however they happen to feel at any given moment. When some kind of change is announced, it's a safe bet that 90% of the response will boil down to "I hate change" with creative variations. As such, the actual comments themselves are unnecessary; the only interesting information is whether the reactions are significantly more or less than the expected 90% of the total volume.

    The signals for "this response is not knee-jerk reaction" are things like: longer than average (but not too long), tidy paragraphs instead of sprawling run-on sentences, maybe some formatting like lists and headings, basically anything that indicates the author spent time thinking about the reader instead of scrawling the hottest take they could. None of those things require close reading or empathetic consideration. It's possible some useful and interesting responses get accidentally scooped up with the unhelpful ones, but nothing is ever perfect.

    As for deleting or at least hiding such comments, if someone is serious about listening to interesting feedback, they'll want to come back and re-read that feedback, and show it to other people. That's a lot easier without the unhelpful distractions, even if only because they don't have to scroll so far.

    The point is that there's a conflict between the idea of theming, which requires a rigidly-defined list of standard widgets, and the idea that application authors should be allowed to customise widgets or invent new ones if they think it's appropriate. The design of GNOME's desktop (or any desktop) or GNOME's default theme (or any theme) or GNOME's theme API (or any theme API) are not relevant to the discussion, because this conflict has affected every desktop based on every GUI toolkit.

    It is, though. If GNOME would have a default theme designed by, say, God, that everyone found completely perfect there would be no need to have theming support at all.If the default theme was something like the default iOS or Mac OS X theme, people would find it acceptable. And if the default theme was butt-ugly, it would be imperative to have theming support.

    It's literally impossible to have a theme so perfect that everybody likes it; the best you can hope for is a theme good enough that changing it is more trouble than it's worth. I'm pretty sure it's also impossible to have a theme so ugly that everybody hates it; Amiga Workbench 1.0 and Windows 1.0 didn't immediately doom their respective product lines.

    A widget toolkit with a theme engine makes some things awesome and some things terrible; a widget toolkit that can't easily be themed (like macOS and Windows have) keeps everything mediocre. Is that better? Maybe? Sometimes?

    I'm guessing that the distros' default themes aren't exactly crazy, just making shoddy applications look a bit wonkier than on the dev's machine. On sane DEs like Xfce, I have zero issues of this kind, which makes me think the real issue is that they don't want users to be able to "damage" their precious branding.

    More likely, Xfce's default theme is close enough to boring grey-and-blue Adwaita that it doesn't cause any problems. Some Linux vendors have much bolder branding, say in brown and orange, and just to stand out they do unusual things like have menus and toolbars in light-on-dark while the rest of the application is dark-on-light.

    GTK+3's theme engine I guess doesn't make it easy to have wildly different colour-schemes for different parts of the window, so themes that want to do that have to add a bunch of custom rules to tweak each supported application... but there's no way to limit a particular rule to a particular application. So if a standard GNOME app happens to name one of its toolbar widgets "switcher", the Linux vendor theme says "widgets named 'switcher' use light-coloured text". Then a third-party app happens to also use the name "switcher" for some other widget that's *not* in the toolbar, so it winds up with light text on a light background.

    > There's absolutely no need to start whining about this, when you could just not implement the feature if it breaks your obscure widgets.

    You do realise that application authors and GTK's maintainers are different people, right?

    > Unicode was a mistake, and it just keeps getting worse and worse (emoji, RTL, Chinese characters breaking BMP). The Europeans already had ISO 8859-1, the Russians KOI8-R, and for the web there was HTML entities. Since everyone who uses the Internet speaks English anyway, computer localization is just a waste of time and money and unpleasant for generally everybody. It would be far easier if you knew that everything everywhere would be en-US, no surprises.

    To go from "I do not need this thing" to "None of the six billion humans on the planet needs this thing" is quite a feat of extrapolation. Does it genuinely not occur to you that if you don't understand a thing, maybe you just haven't yet encountered the problem it solves?

    > In fact, I would prefer it if they picked a language completely at random, even if it weren't English, and replaced the numeric warnings with that. At least you can memorize the meaning of _CRT_SALAMA_HAKUNA_ONYO, and you already got one of the three words down if you watched the Lion King as a child, which is more than you can say for 4996.

    Not strictly relevant to the point under discussion, but I do want to point out PHP's famous "expected T_PAAMAYIM_NEKUDOTAYIM" error.

    > It's a damn shame there isn't any good open-source forum software.

    GNU Mailman and INN say hi.
    Kawaoneechan That's almost the ABXD origin story.
    CaptainJistuce
    Posted by Kawa
    Imagine, an ABXD with sharting sharding. What even is that?

    That's when a quarter of the community leaves to post on a shard board started by someone with an axe to grind, right?
    ‮strfry("emanresu")
    Posted by tomman
    So all message boards must scale because they COULD have as many users as Fecesbook or Reddit someday. That's an... interesting rationale. And by "interesting", I mean completely nuts.

    As much as PHP is the fractal of bad design, there is nothing wrong with message board platforms like ABXD, phpBB or SMF for small boards like this one, or somewhat bigger boards with a couple hundred active users shitposting every hour.

    Not everything has to be BIG or fancy, y'know. It's like quoting a customer for a Cray supercomputer if all they need is something to host their HR/payroll app for their 3 users making paychecks for 250 employees. An supermaket Lenovo/HP shitbox will do the job just fine.

    You're right in that this would be a completely insane line of reasoning, but I have never said that.
    Posted by sureanem
    It definitely is the best open-source forum software for small boards that don't need those features however, insofar as it is open-source.

    Posted by sureanem
    Stuff like thread subscriptions/quote notification, guest posting, image uploads, thread previews, and such. Without those, it just devolves into complete chaos and a shouting match given enough users. Reddit is a great example of how this eventually happens.

    92 users is far from enough. Even 9200 won't do the trick. I'm talking ~1 million or more. One post per minute or more, something like that. For such forums, there is no good FOSS software.
    Kawaoneechan Imagine, an ABXD with sharting sharding. What even is that?
    tomman So all message boards must scale because they COULD have as many users as Fecesbook or Reddit someday. That's an... interesting rationale. And by "interesting", I mean completely nuts.

    As much as PHP is the fractal of bad design, there is nothing wrong with message board platforms like ABXD, phpBB or SMF for small boards like this one, or somewhat bigger boards with a couple hundred active users shitposting every hour.

    Not everything has to be BIG or fancy, y'know. It's like quoting a customer for a Cray supercomputer if all they need is something to host their HR/payroll app for their 3 users making paychecks for 250 employees. An supermaket Lenovo/HP shitbox will do the job just fine.
    Kawaoneechan Cos I like it.
    ‮strfry("emanresu") Why C#?
      Main » Discussion » GNOME: "Please don't theme our apps" » New reply
      This does not actually go there and I regret nothing.