Date: Sat, 10 Aug 2019 17:56:47 +0200 From: hw <hw@adminart.net> To: Polytropon <freebsd@edvax.de> Cc: freebsd-questions@freebsd.org Subject: Re: thunderbird in German? Message-ID: <874l2p9fpc.fsf@toy.adminart.net> In-Reply-To: <20190810133516.3254f6e5.freebsd@edvax.de> (Polytropon's message of "Sat, 10 Aug 2019 13:35:16 %2B0200") References: <87o91bw4mx.fsf@toy.adminart.net> <20190731012754.3783a859.freebsd@edvax.de> <878ssc5v3x.fsf@toy.adminart.net> <20190802053434.6babd0d0.freebsd@edvax.de> <87y30b8h0v.fsf@toy.adminart.net> <20190803042205.550f5903.freebsd@edvax.de> <87pnllbxp7.fsf@toy.adminart.net> <20190805070522.be1fb873.freebsd@edvax.de> <87sgqdq0nw.fsf@toy.adminart.net> <20190809133419.3de9e65f.freebsd@edvax.de> <87pnlevzbb.fsf@toy.adminart.net> <20190810133516.3254f6e5.freebsd@edvax.de>
next in thread | previous in thread | raw e-mail | index | archive | help
Polytropon <freebsd@edvax.de> writes: > On Fri, 09 Aug 2019 22:50:00 +0200, hw wrote: >> Polytropon <freebsd@edvax.de> writes: >>=20 >> > On Wed, 07 Aug 2019 02:28:19 +0200, hw wrote: >> >> Polytropon <freebsd@edvax.de> writes: >> >>=20 >> >> > On Sun, 04 Aug 2019 02:07:00 +0200, hw wrote: >> >> >> Polytropon <freebsd@edvax.de> writes: >> >> >>=20 >> >> >> > On Fri, 02 Aug 2019 22:10:56 +0200, hw wrote: >> >> >> >> Polytropon <freebsd@edvax.de> writes: >> >> >> >>=20 >> >> >> >> > On Fri, 02 Aug 2019 01:22:42 +0200, hw wrote: >> >> >> >> >> Polytropon <freebsd@edvax.de> writes: >> >> >> >> >>=20 >> >> >> >> [...] >> >> >> > While it has been good practice for decades to use -Wall and its >> >> >> > equivalents, it doesn't seem to be important anymore. Compile t= ime >> >> >> > errors and warnings are irrelevant just as runtime warnings and >> >> >> > errors. >> >> >>=20 >> >> >> My impression is that software is now generally much more stable t= han it >> >> >> used to be. I don't know why, perhaps better tools became availab= le. >> >> > >> >> > Yes, program stability has improved, especially for web browsers. >> >> > But it's more than fair to acknowledge that the complexity of a >> >> > web browser is comparable to the complexity of a whole operating >> >> > system. >> >>=20 >> >> Web browsers are evil. >> > >> > A neccessary one, it seems... >>=20 >> It's not necessary; only the functionality needs to be kept within >> reasonable limits. > > Are there limits? If yes, where? > > You can see exaggeration regarding limits in some web pages: > They want to provide functions the web browser already delivers, > like bookmarking, printing, fonts & sizes, and the "back button" > of course. Is this needed? Probably not. Is it done? Of course. Is it reasonable for web browsers to turn into front ends for so-called web applications and into programs that can do virtually everything, like even display images, play videos and download files? With a couple tabs open, Firefox occupies 5GB of RAM and creates significant CPU load all for nothting because I'm currently writing an email and am not using it. This is not reasonable but ridiculous. Why doesn't at least Ctrl+z work? The purpose of a web browser is to make information accessible by displaying text which may contain references to other text that can be followed. That's all, and it's all I'm interested in when it comes to a web browser. I would consider that a reasonable limit. For other things like displaying images, playing videos, managing downloads, providing a front end for web applications and whatever else is useful, there should be other programs that may be much better at what they are doing. > [...] >> >> I don't know. Did Gimp crash or something? >> > >> > No, it works as intended, no problems. Just those messages. >>=20 >> Maybe then there isn't a problem --- and/or it means that Gimp handles >> broken files gracefully. > > But the error message isn't any helpful! That goes for a lot of, if not most or all, error messages. "Press any key to continue" is basically an error message. What does it tell you? There is no "any key" on my keyboard and the mouse buttons have no effect, so what am I supposed to do? Why can the computer not continue --- whatever that means --- all by itself? After all, that's what I have it for, isn't it? > What does "(gimp:3022): GLib-WARNING **: goption.c:2132: ignoring > no-arg, optional-arg or filename flags (8) on option of type 0" want > to tell me? What have I done wrong, what should I do instead? Or about > what does the program want to inform me? I have no idea... in my > opinion, those messages belong into a "debugging enabled" mode. You still wouldn't know what they are supposed to tell you. Why doesn't your car give you an error message but only turns on a warning light? If you knew what's wrong with it, you might easily be able to fix. Live could be a lot easier if people were more reasonable and would give up insisting on making themselves helpless. Instead, they have a bad need to celebreate unreasonability (or: the unreasonable) and to encumber themselves with it. Maybe it's a desease just like religiousness. OTOH, what is the programmer supposed to do when creating an error message? Users nowadays are incapable of reading anything, especially error messages, and they are way too incompetent to understand error messages anyway. Omitting all error messages or changing them to only say "error" is a bad idea because that makes it so that the programmer can't figure out what went wrong and hence can't fix it. Providing error messages leads to users complaining about getting elaborate error messages (and that they don't understand them) in some cases and questions about something in others where the error message didn't sufficiently elaborate. Denying an operation when the right thing to do would be to provide an error message so that a problem affecting the operation can be fixed before the operation is performed is not always an option, either. Providing error messages that allow the programmer to fix a problem only in debugging mode isn't a good option because the error will occur while the debugging mode is disabled, and the user encountering the error does not know how to reproduce it because they never know what they are doing anyway, so the problem can't be fixed even assuming that the user would tell the programmer about it, which they usually don't. Even when there are no bugs in the software, users still make mistakes or suddenly want to do things in a different way, and error messages can be required. Do you have a good solution for this? >> > Maybe I should decorate my programs with such warnings, too? ;-) >>=20 >> As long as they work gracefully even under adverse conditions, nobody >> would mind. > > ***warning***critical***: program: File foo.txt [0], index 4. > Yes, this definitely helps. ;-) Yes, it does because you can fix the problem or discover that there is no problem or find out that you need to modify your program or that the user needs to be educated or that you can remove the message or do whatever. You wouldn't know without. >> >> I don't want all the windows to look the same, that's ugly and boring, >> >> and some need to be handled differently than others for things to be >> >> usable. >> > >> > Yes, because even though programs have common ways for inter- >> > action (the known GUI elements), they serve different purposes >> > and therefore it's _significant_ to provide an interaction that >> > matches the task you want to perform with the program. >>=20 >> That doesn't mean everything has to look the same and needs to be >> boring. > > If you compare web pages, they strangely "all" look the same, > more or less, That's because you're displaying them all using the same web browser. > and most of them are boring. You can always look at other web pages. > "Flat design" which is currently modern requires lots of screen real > estate, to provide what could fit nicely on a 80x25 text screen... ;-) Yeah that really sucks. Most web pages look messed up because I have the web browser enforce a minimum font size such that I am able to read the texts on web pages. See "reasonable limits" above ... >> > Example (seen a few days ago): An accounting program, data entry: >> > You need to click on every input field in order to enter the >> > data with the beyboard. >>=20 >> That sucks. >>=20 >> > Some fields have a button where you can open a "calculator-like" >> > screen keyboard for numbers. But for letters, there's none. You cannot >> > use TAB or cursor keys or ENTER to advance fields. Can you imagine how >> > annoying it must be to enter pages and pages of information to that >> > stupid program? Oh - and you cannot feed external data... >>=20 >> I would refuse to use such a crap. > > Not possible, because it's required by regulations that you > use _that_ program for the specific purpose, because only > _that_ program exists and is certified. Oh, and of course > you have to pay for the privilege to use it. It also does > not remember your 4-part access credentials, you need to > type them in for every session. I'd still refuse to use it. >> >> > A program's job is _not_ to do what the window manager does. A prog= ram >> >> > can request to be handled in a different way, for exaple, without >> >> > a window decoration (like XMMS MP3 player where it doesn't make sen= se, >> >> > or little utilities in the "system corner" like xconsole, xbiff, >> >> > xclock, xload, xcpufreq, etc.). But programs sometimes used their >> >> > own ugly white _mouse cursor_ for no reason! Using a specific >> >> > mouse cursor is normal for tools like Gimp, xfig, even LibreOffice. >> >> > But generic programs? Why? >> >>=20 >> >> Like emacs using a particular pointer to indicate that it is busy when >> >> configured to do so can make sense. If it couldn't do that, it would >> >> need to somehow tell the WM that it's busy and the WM would have to >> >> change the pointer on some or all windows --- and if it would be done >> >> like that rather than emacs changing the pointer, people would say the >> >> WM and emacs are bloated. Or are they, or is X11 bloated for allowing >> >> to change pointers? >> > >> > If the cursor changes for a good reason, it's acceptable, >>=20 >> Who is to decide what is a good reason and what isn't? > > A good question. Common sense, logic, established standards and > generally accepted procedures should be applied, but of course > not with a stupid strictness that stops developers from creating > new and good features. Isn't that what has happened? >> I have spent so much time waiting on computers that I don't want a >> pointer or anything to change to an icon which indicates that the >> computer is busy with something. I want that something done and not a >> busy computer. Even that emacs is single threaded is very annoying at >> times ... > > Computers are fast and powerful, but programs consume more > resources. The quotient > > resources provided by computer > ------------------------------ =3D const. > resources consumed by program > > seems to apply. No, it doesn't. I'm waiting more because of awfully low internet bandwidth than anything else. Fortunately, we can finally have computers that do not have us wait on them more than anything else. > Computers get faster, programs get more bloated, result of "speed of Bad programming is always an issue. > use" stays the same. Sure, you cannot say that everywhere, for > example, media encoders are much more pleasant to use than 10 or 15 > years ago, but just have a look at the load times of "modern" web > pages. Compare that to "old" web pages on old computers. Notice > something? It didn't get significantly better... Do you have an example? Much of what contributes to how long it takes to load --- and what about displaying it? --- a web page with some computer has nothing to do with the very page itself. >> > but what about a program changing my normal black mouse cursot to an >> > ugly white one? It's a PDF viewer, nothing more! It doesn't switch to >> > any other cursor shape (like Gimp: cross cursor for selection, paint >> > can cursor, brush cursor and so on). >>=20 >> Perhaps who decided it should change has got lots of black PDFs and >> needs this to indicate that the PDF viewer has the focus. > > I'll ask Adobe if their developers open so many black PDFs. :-) ok That Gimp changes the pointer is annoying. Yesterday I had to figure out where the pointer seems to be pointing at because Gimp had changed it and I could only guess where it might point at. Gimp doesn't need to change the pointer because I know which tool I have selected and the pointer doesn't tell me that anyway. I use text, not icons. Icons look all the same. Seriously, is there a way to turn that off? I would. >> It's like the retards who are sending HTML and think it's an email, [...] > > There are lots of real companies that you can pay for designing > you "HTML e-mails", usually for presenting ads and tracking the > users. They automatically go into my spam folder. If someone wants to pay someone to create emails like that, they should also pay me for the waste of my resources. Besides, advertising in public should be illegal. I don't want to be bothered by it, and it has nothing but bad effects. >> [...] or >> web pages trying to use their own fonts as if the idiot designing the >> page had any idea what it might look like on my screen and what fonts I >> like and which ones not. Who do they think they are? > > They are designers and developers. Some of them even have a few > typography skills. They assume the web as a pixel-perfect medium, > and they design for it as if it was a brochure or billboard ad. > Only a narrow subset of devices will display the web page the > same way as the designer sees and intends it. But as long as > he can present something in the meeting that pleases the PHB, > his job is safe. For now. :-) They need to realise that it is entirely up to the beholder how some HTML page is being rendered and that it is irrelevant what they think a font that looks good and is readable. > The Opera web browser has a handy button I call "Blatt / M=C3=A4nnchen" > (paper / man) where you can switch from what the designers coded > in HTML and CSS to _nothing_ (browser defaults). This button makes > a lot of web pages readable! Sounds like I should try it out. >> That you can't protect anyone against their own stupidity doesn't mean >> you should support it, unless you're a big meanie. > > In my opinion, you can try to help those who want to escape the > vicious cycle of stupidity and consumption, but it will work only > in very few cases, I fear... No, you can't. Try to help someone and you will be a looser. Even if someone is asking for help, they usually don't want any and will turn impolite, try to insult you and perhaps call you a troll. That doesn't have anything to do with stupidity, it's just the way it is. I think this used to be all different and am wondering what changed. > [...] >> > Elections miss a "none of those", which is problematic especially >> > in Germany where shit is offered in different tastes; interestingly, >> > if I remember correctly, India (!) has a "none of those" option... >>=20 >> That isn't enough. The only votes possible should be against someone or >> something, and what- or whoever has the least votes against it or them >> wins. > > This is something I have in mind for decades: Imagine a ballot > where you can only "unvote". You select the parties or candidates > that you feel won't be good for the job. At a certain "aversion > rate", a party has to be dissolved because it has no backing in > the society - because voters have shown (!) the opposite. Combine > this with personal (!) responsibility where it makes sense, and > you will see that only those who _can_ do a certain job will be > elected into that job. Maybe ... I wouldn't require parties to dissolve themselves and only cut their funding. And then, why shouldn't parties be illegal? What's the point of parties? Voters can never know "a party" as they could know a person, and why should a voter be requested or allowed to vote for or against anyone they do not know. Voters should be required to prove that they are sufficiently informed and reasonably capable of voting. >> >> >> But I haven't figured out how to make it so that libreoffice inste= ad of >> >> >> gedit is the default program to open spreadsheets attached to >> >> >> emails. >> >> > >> >> > Right-click on attachment, select "Open with", enter the command >> >> > needed for this file type, and it will be saved. >> >>=20 >> >> It saves only the command and not when to use it. That is way too >> >> complicated for the users because they need not only to remember to p= ick >> >> it but also which one. >> > >> > I have learned to live with it, but I find it annoying, too. >>=20 >> Hm, that's really bad. Maybe I should make a feature request. > > Do that. It would be a good feature, because the action "double- > click on attachment" currently does nothing, whereas the action > "open with" can be selected from the context menu if needed, > just like "save" or "save all". I doubt it would be considered. Why bother. >> >> [...] >> >> Why can't there be a button at the attachment I can click on to open >> >> the attachment with the program I once, and only once, picked for it? >> > >> > This is the concept of the common file associations in file >> > managers. Sylpheed could pick that idea up. >>=20 >> File managers don't have a button like that. > > Not a button, but if you double-click on a file icon, the file > will be opened with the associated program; if no such program > is defined, a dialog will request one, with the option to make > the choice permanent. Yes, they don't have a button like that. >> Then why is the HTML junk optionally displayed? That isn't an email and >> nothing I would want to see. > > Sylpheed does not render HTML. In a case where the email contains > additional HTML, it's presented as an attachment. If you click > on it, it looks the same like the text version because (as I said) > HTML will not be rendered. If _only_ HTML is present (which > violates the standard), it will be displayed like text. Replying > will _always_ use text, as intended. Ah I thought it would optionally be rendered. I like that it isn't. Who says it violates a standard when an email has attachments and no text, and which standard is that? That could be good to know. >> >> > For example, I have two different commands for PDF attachments - >> >> > simply because I sometimes need program A, sometimes program B, >> >> > depending on the PDF file. >> >>=20 >> >> Users don't have that. >> > >> > Exactly. That's why Thunderbird does a better job in this case. >> > It "just works". Have a PDF attachment - doubleclick - system's >> > PDF viewer opens it. >>=20 >> Where do you live? It doesn't work that way, especially not "just". > > After installation of the required programs (!), at least on > "Windows" it tends to work that way. No, it doesn't. If you install /one/ PDF viewer, you still have to tell others programs to use it. > On the Linux desktop, xdg-open, using MIME types, is the preferred way > of doing things, according to the X Desktop Group. But who is using that? > So it should apply to FreeBSD as well, but not all desktops and of > course not all programs seem to support it. This, in my opinion, is > because all the desktops do so many things differently that a > consensus is hard to reach. A consensus is hard to reach because humans have become particularly bad at working together. Perhaps they always were and only necessity forced them for otherwise they would have become extinct. Besides that, a lot of things are being done because they can be done, and someone likes something better than something else. I think giving choices is better than taking them away, and the so-called "desktop environments" are designed to take choices away. People can't be bothered to make choices. Only _if_ they realise they don't have any left _and_ suffer from something, they start trying to make them and things suddenly get nasty once there are enough of them.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?874l2p9fpc.fsf>