That was me, and its not Dubious.

Yeah it is. You claimed that if you went to RedHat with huge wads of cash they still wouldn't do it. Unless you've tried, I doubt the veraity of that claim. I don't require proof: I'd take your word for it.

You can find a consultant to support it, Im sure, if thats any consolation.

That was my point. Being OSS, you're not beholden to the original vendor so it's less important that RedHat support it. If you really need, you can still get someone to support it.

Arguing that Microsoft is "bad" because theyre not FOSS (which is really what you are driving at)

False premise: I made no argument about microsoft.

  The OP claimed that you couldn't get support on very old OSS software from the vendor (though this in itself is a dubious claim). I pointed out that because it's OSS, you can in fact offer anyone great wads of cash to get support because you're not dependent on the vendor.

Note how I make no claims positive or negative about microsoft. There's not even an implied claim, since apparently if you pay enough, MS will continue to offer support well past even the last final honestly we really mean it now deadline.

Ideological spiels about how Windows

You know what's worse than ideological spiels? Hallucinating the existence of idealogical spiels in the words of some hapless slashdot poster, where no such thing exists.

they are choosing the greater of two evils.

No.

Eventually supporting too many screwy and ancient systems starts to cause just so many problems that it is really, really hard to write solid, well tested, clear code. The heartbleed bug was exactly a result of this. Because of supporting so many screwy platforms, they couldn't even rely on having malloc() work well. That means they had their own malloc implementation working from internal memory pools. Had they not, they would have benefited from the modern mmap() based implementations and you'd have got a segfault rather than a dump of the process memory.

Supporting especially really old systems means having reimplementations of things which ought to be outside the scope of OpenSSL. Then you have to decide whether to always use the reimplementation or switch on demand between the custom one and the system one and whether or not to have some sort of quirk/bug correction.

This sort of stuff starts to add up and lead to a maintainance nightmare.

What OpenBSD are doing: throwing out all the accumulated crud and keeping the good parts is a fine way to proceed. It will almost certainly be portable to the other BSDs, OSX and Linux since they provide similar low level facilities. I doubt a port back to Windows would be hard because modern windows provides enough sane facilities that it's generally not too bad for heavily algorithmic code like this.

Basically there's no way for them to get started except to first rationalise the code base and then audit it.

According to the UN, Taiwan is part of China.

Yes, but there is no meaningful definition by which Taiwan is part of China. It is politically and economically separate in every regard, and the people living in it regard it as not China. They have different laws (PRC laws have no bearing in Taiwan), a different government and an independent military. It's also got a bunch of pretend-not-embassies from quite a large number of countries including the US.

The only reason the UN don't recognise it is because China has sufficient influence to make the member states maintain a rather silly fiction.

The thing is, you can still call MS and get them to support XP if you beg with enough money. Good luck calling RedHat and getting them to support RedHat Linux 7 (not the enterprise one, the 2002 one).

Actually I'd bet if you offered enough money you could get them to support it. But you know what? It doesn't matter if they refuse because you have the source and can pay your own person to support it if you are really desperate.

That's the nice thing about open source. You aren't dependent on the original vendor. Hell you could get someone to support any OSS if you begged with enough money whether or not the original vendor is still in business.

Even the ones that weren't illegal at all and that were doing a lot of good. Either you are being deliberately ignorant or...well I can't really see it any other way.

Well, how about you enlighten us with a [citation]. So far we know the NSA were spying on their own citizens, illegally spying on their allies and were being helped by GCHQ in yet more illegal spying. The only piece of vaguely positive news I heard from the whole thing was that at least GCHQ were charging the NSA for their nefarious activities, rather than doing it for free.

I think you misunderstand the role of a Window Manager in the X11 stack. It's a very specific part of the stack with a very limited role. It take no part in most of what goes on.

The problem you hare having is you have the wrong definition of GUI in your head where you are using it to mean using a window manager.

I disagree. I use GUI to mean the whole system. The WM is one very small part of that system.

A GUI has a bunch of features it needs to achieve: windowing system (window manager being an example),

No: a WM is not a windowing system. X11 is the windowing system. A window manager is simply a program runing on top of X11 which moves windows around.

menuing system, interprocess communications, widget libraries.... Running a

The menuing system is provided by the widget libraries (gkt, QT, etc). The IPC is provided by X11 for things GUI related.

window manager to manager terminals doesn't mean you are using a GUI you are just using window management.

Well, perhaps. In the erly days, that would suffice, maybe now it doesn't. Nonetheless I use more than terminals. I like terminals, but I use the GUI driver for gvim, run web browsers, sometimes Eagle Cad (most definitely a GUI), sometimes LibreOffice, the GIMP, Inkscape, etc. There's also other bits and bobs, like a notification area holding icons for messaging clients (pidgn, xchat, skype), a CPU, battery and temperature monitor, some bluetooth widget and so on.

When I said "you aren't even using a GUI". I wasn't being insulting or closed minded or anything else. I was simply saying the graphical setup you are using does not contain features required to classify it as a GUI.You are essentially in the pre GUI age where you have graphics on screen.

Fair enough, but I do believe you are mistaken however.

What do you think the GStreamer library does for Gnome? Or Kstreamer does for KDELIbs. No those components are not well separated, they are meshed together where applications are just wrappers around libraries.

They play videos? I'm not sure I understand the point. The WM itself takes no part in that role. Anything GStreamer, Kstreamer, FFMPEG etc based works just fine without even a window manager running.

You are right about Terminology. That is substantially more advanced something like 2005 in terms of OSX and clearly quite advanced for Linux datatypes. I'd have to know more but I'd agree that isn't 90s technology.

I don't think OSX has any equivalent, though I haven't been following the OSX thing. Terminology starts to seriously blur the line between the commandline and the terminal. I'm deeply impressed.

As far as 1994, in 1994 I was on 64 bit Suns switching over from SunOS to Solaris. I also used AIX sometimes. I hadn't been on 8 bit in a dozen years.

Fair enough. I had a sumer job in '94. I got to use an HP (running CDE) and occasionally an SGI. 8 bits were obsolete then, but that's mostly why I could afford one of my own.

I started using Linux in 1995 on cheap home computers and FVWM was the window manager that was most popular then.

I started a few years later with Redhat 5.2. The situation was similar.

Your setup is from that time. Look at Caldera Open Desktop or RedHat from the time periods. By 1999 KDE is mature enough as a GUI that people are building whole systems around it. More or less if you aren't using an integrated GUI you are pre-1999 i.e. 1990s type system.

I'm not sure I agree. Some of the components date to then. The styling certainly does, but a lot of the core, most of it I'd argue is more modern. I mean I could use GnomeShell and GnomeTerminal, but there wouldn't be any more integration than I currently have. As far as I can see, there's no integration missing that I'd get if I was running gnome.

The X11, freedesktop.org people have created and collated a very nice set of standards which allow X11 based programs to interoperate. This is why KDE programs work perfectly with a Gnome based WM running and why they both work fine under my system. The ICCCM, XDnD protocol, tray protocol, notification protocol etc describe these things.

Mixed paradigm languages would be things like Scala or Clojure.

Ah fair enough. I was under the impression that things like Ruby and Python allowed some degree of functional type programming.

Tizen is the full GUI/OS that includes Enlightenment. What would change is all your applications and your window manager would all be using EFL so functionality like messaging and notifications could pass between layers effectively. I was saying that Terminology is not a GUI, even Enlightenment is not a GUI but Tizen does have a GUI because it layers everything on top of EFL.

Sounds neat, though in X11 land those things (notifications etc) all work across libraries because they're defined as protocols rather than API calls.

Perhaps you have missed it, but in our last release we released a preview release on FreeBSD?

I had missed it. That's good to hear, to be honest. Excellent news, in fact. In that case, I withdraw my criticism over that.

I'm still very un-fond of some behaviour of the file dialog box. If you're around to listen, I'll happily elaborate.

I'm not trying to dimiss people "who hate change".

Well, that's actually fair enough: some people are going to hate change no matter how beneficial. The problem is dismissing people who have thoughtfully rejected some of the recent changes as merely hating change in general.

I find myself in the latter camp, but often dismissed as hating changin.

But you can see how GNOME discussions have transitioned to Systemd,

Well, yes. systemd has been greeted with suspicion. I'm not surprised: there's a lot of FUD on both sides. One would expect it from the proponents of the system being displaced but there's been plenty of pro systemd FUD. There are also some parts of the design many people find rather dubious. I don't know. My new systems have it and I can't see any real difference, but I've not looked in detail.

The thing that has annoyed people about GNOME is that they've pretty much forced the issue on systemd. Since it's a dependency of GNOME and GNOME is the most popular environment, most distros have switched pretty much without regard to the merits. It seems odd that a desktop environment needs such deep vertical integration as KDE, LXDE and XFCE do not and are still featureful.

What I think is that people deeply resent GNOME essentially forcing them to adopt another system they dislike.

But ultimately, people are upset because we aren't adhering to status quo.

No, I think you're very fundementally misunderstanding what people are objecting to. Many people aren't objecting that the status quo is changing, they're objecting that it's changing in ways they don't like. It's very, very easy to dismiss those people simply as "hating change". I think that's a mistake to do so.

If you've been arguing like this and I'm tlaking about everyone from kernel developer to random people on the internet, you can start seeing the patterns.

Indeed, but if you've been around long enough you also start to recognise the cascade of attention defecit teenagers model. That involves generally change for the sake of it by people who don't understand the original system enough to avoid the mistakes it made and the mistakes it didn't make.

TL;DR don't confuse people disliking your changes as people disliking change in general.

OK but you would agree that with your profile i.e. using a mid 1990s setup with a few slight advances, you fall pretty squarely in what most people would call the "hating change" camp.

OK yes, but I'd say that was ignorance from those people and is every bit as bad as those that do genuinely hate change. I don't hate change: I embrace it when it benefits me. I try to avoid changing things for the worse. I'm not a fan of churn for its own sake.

This I think is the problem with quite a few people at the forefront of change. They are so invested in it that they assume everyone who doesn't like what they've done must be a luddite nd therefore has an invalid opinion which is best ignored. In that case they're all too eagar to dismiss other opinions.

  Obviously you have been around long enough to deal with change better than the people freaking out about Gnome, you mostly ignore it, but I certainly don't think of you given that description as a change enthusiast or anything.

Perhaps not an enthusiast. I like new tech and seek it out. I don't however adopt wholsale change merely for its own sake. Some things I've fond were done very very well a long time ago and it's going to take something quite exceptional to improve on them.

As far as C++/BTRFS those would be examples where you clearly are a change enthusiast. A different areas of computing.

Again, I'm not sure enthusiast is right. The new things are for me just better. BTRFS is much less faffy than LVM, so makes multi-disk stuff a breeze. C++11 is just plain great. Actually, I feel the C++11 committee are quite close to me philosohically. They do adopt new stuff, but only when it really does help and try not to break old stuff without a really good reason.

Well the big changes in terminals for English speakers are transparency and tabs IMHO.

Two things I can leave, TBH. I did download one of the early terminals with transparency support then went and found a really cool desktop pic (a martian frying the thunderchild iron clad---yes it was the 90s and yes I was a teenager then) and set it up so it looked just so and (to my mind) super awesome. Then I tried to code and realised that a plain black background was actually superior. :)

It is mainly with other languages

Yeah fair enough. Though even the venerable XTerm and Fixed Semi-Condensed font are now far better in their unicode support than previously. I wouldn't know correctly rendered Hindi if it ran up and bit me on the leg, however.

Terminology, is just an Enlightenment app from the 1990s.

The first release was August 2012. It's the terminal that allows embedding of images and videos etc within the text amongst other things. It's a real GUI terminal.

  Were you using Tizen I'd see more of a move towards a GUI and away from an admittedly cool windows manger. But you aren't even using Enlightenment across the board.

Nope, I'm using FVWM since I like the way I can set things up. I'm only dimly aware of Tizen. What would it change?

Ah now I see what upset you. A windows manager is a component of a GUI but a window manager is just a small fraction of a GUI. The widget set and the interaction subsystem (event handling) are mostly not part of window managers.

I'd say barely at all. Almost all events go straight to the program in question. A very few get bounced via the window manager (basically only ones to do with window placement) so that it can draw borders etc. The compositor now eats a few extra positional based ones so it can deal with funny window placement. The WM itself has no influence on the main path of most events, or the widget set. Some WMs don't even use any widgets at all.

This is the debate that happened with KDE 1, whether wanted a GUI or just wanted fully featured window managers. If I were to ask you "what facilities does FVWM2 provide for database access over a network?" or "how does FVWM queue QoS video streams vs. non-QoS video streams?" you get an idea of what FVWM2 doesn't do that means that it isn't a GUI at all. This isn't zealotry but rather the very definition of the word "GUI".

I'm not really following to be honest. Surely the network susbystem of the kernel or router deals with QoS. The only thing the WM does is deal with placement of windows on the screen. The compositor on Wayland fills a very similar task. The components are well separated and none of the WMs including the GNOME and KDE ones have any part in playing video streams.

Exactly. FVWM2 has no idea how cut and paste works. Object communication is what GUIs have to do.

But copy/paste is handled by the X server (along with DnD). The mechanism is well specified and it's quite easy to implement. The window manager is purely a device for arranging windows on the screen. That's one of the nice parts of X is that the WM can be switched out easily and none of the mechanisms are affected. In fact it's possible to run X with no WM (unpleasant but possible), and copy/paste still works.

But if the tools actually make use of graphical objects that falls apart.

Well, to some extent yes, depending on what facilities are on offer.

Cut and paste being a perfect example of where reducing everything to commandline fails terribly.

Well, copy/paste commands which extract text from the X11 clipboard and echo it do exist (I've even written one of my own), but I've never had much use for them to be honest. Copy/paste generally stays in GUI land for me. Though I mostly use it within a single editor instance.

In all fairness. C++ is a rather traditional language. GCC is a rather traditional system... Using gvim rather than a IDE to do C++ programming is super traditional.

Kinda. C++ still has ideas in it that most other langages have yet to adopt. Additionally, the new versions have all the modernish goodies like type inference, lambdas and so on. It's a system with a long history, to be sure, but then so is the Linux kernel. From a practical point of view, modernidiomatic C++11 is almost unrecognisable compared to what was state of the art 20 years ago.

Even vim itself is a long way different from the vi clones of years gone by. You can get a LLVM plugin now for instance which does fully integrated C++ autocomplete for example, something traditionally which is an IDE feature.

Mixed paradigm dynamic languages with libraries that tie them to web and database (i.e. 2010s equivalent of 4GLs) are newish.

Only vaguely. Ignoring the libraries part, I remember tinkering with Python in the 90s. And TCL fits the dynmic, mixed paradigm bill fine, even if it is in most ways truly hideous. As for libraries, well, PERL was there first. I'm not doing down the new ones, PERL is also vile, but the ideas and even the languages are not that fundementally new. That said, if I was going to be doing some sort of web service development, I'd probably reach for one of the newish dynamic languages with loits of libraries.

If you are talking performance then languages that take better advantage of today's processors than C++ make more sense.

Mostly I do image and data processing. Performance is generally important, but also C++ is about the best development environment for the type of thing I'm doing. I've tried matlab, octave, numpy/scipy and others and I keep coming back to C++.

Look at your setup and ask yourself what couldn't you have been doing 20 years ago? What fundamentally is different about what you are doing and what you would have been doing 20 years ago?

20 years ago was 1994. I was still using a BBC Master then (checking on wikipedia, they were discontinued in 1994 so I guess they were available cheap which is why I had one). If you're not British, then filling you in, that's one of the more capable 8 bitters. At work I used some HP thingy running CDE. Oh and the SGI. Oh the sgi. Seeing that demo stuff in 1994 was -mind blowing-.

But I'm not sure what you mean. Then as now, I spent most of my time typing code into some sort of editor. But the sort of things I can do now are much more advanced. The quality of languages has improved so I can write bigger, more complex things single handed. I can collaborate far better than I used to be able to (another thing---I'm a DVCS evangalist now). I regularly use algorithms invented only in the last decade. I can process whole videos, something almost unimaginable 20 years ago. My computer is also my music player as well, which is nice.

If you're asking if I would give up all the modern features, the answer is no, not a chance. Even with my 1994 era FVWM config (yes really, I started configuring it in 1994 at my first job and have been tweaking it ever since which is why I still have pre Win-95 syle window decorations).

Like I said, I don't hate change when it's for the better. I've been keeping up to date with the latest developments in GUI land, and I'm glad that people finally got around to implementing Copy/Paste as the ICCCM specified all those years ago so we could move more than just plain text around. But in terms of placing windows, FVWM has yet to be beaten.

For other parts of my setup, on the electronics end things are amazingly different. Were now awash with sub 3V devices whith integrated SPI or I2C busses which can be strung together to an insane number of different microcontrollers. This makes tiny low power coin-cell powered devices with in the range of a low-end commercial engineer or even or hobbyist. Though vendor tools have often not improved much since the 90s. But the stuff I can build now and the vendors and the tools available to help are just incredible.

On the hardware end, 3D printers are a complete game changer for me. They are just wonderful machines.

Honestly, I was a little disappointed by your reply. However, I recognise you as someone round here who I generally respect so I'll give you the benefit of the doubt with a proper reply.

How is any of this proof of the reverse?

I'm not sure what you mean. I'm not trying to prove anything. I'm only trying to illustrate that dismissing people who don't like the new gnome things as "hating change" is every bit as bad as the gnome haters dismissing gnome merely because it's different.

You are basically saying you are starting to appreciate the improvements in Linux that too place in the late 1990s.

No, I'm not. If you actually read my post you'd see that. I love the new C++ standards. The most recent one was ratified in 2011 and the current one isn't even complete yet. Doesn't stop me trying out the latest SVN of GCC. Sure it's different, but the change in that case is for the better. Plus the new GCC has many more neat features.

Likewise, changes in the kernel have been great: I run BTRFS on this laptop and that's not even a stable release yet.

I mentioned bluetooth 4. Another recent thing I really like: I specifically purchased a Nexus 4 phone just so I could get started with this new feature quickly. That was only ratified in 2010, with hardware following later. In fact that's the one that got me started on bluetooth. Up until that point bluetooth had not provided me with anything I personally needed. Version 4 has some really compelling new features, so I jumped right on board with not a second thought.

These things are all from the not yet complete 2010-2019 decade so dismissing them as "1990s" stuff is, frankly, silly.

I mentioned the system tray stuff too. The standardised system dates from 2004, not the 90s. You can verify this by looking at the dates on the documentation.

I use udisks happily to mount and unmount pluggable media from the commandline, which is simply the CLI interface to the mechanisms that the more popular GUI file managers use. It's a bit nicer than the old mount based system (though that was hacked to work in the same way), mostly because it seems to be substantially less prone to races. No more /mnt/sdc1 for me. I do much prefer /media/name-of-disk. Again another change I've adopted.

And again, you simply ignored my impending switch to Terminology from XTerm. Yeah, there have been a bunch of terminals (e.g. gnome-terminal), but not a single one of them has actually provided compelling features to me. Mostly it's been change for change's sake. Gnome terminal, for example added AA fonts (I like AA fonts in general, but I happen to like Fixed Semi-Condensed for terminal, so AA adds not much to me), a menu bar (in a terminal that takes up space and adds little useful), tabs (I find the WM capabilities offered by FVWM tend to take priority over tabs for work involving terminals so I don't use them). It allows GUI based switching of profiles (XTerm does offer profiles, but not GUI switching), but I never used them anyway, and background pictures which I certainly don't use. One new very recent feature that GNOME-terminal does offer is text re-flow which is semi-useful. However, that's very very recent and I'm switching to terminology anyway since that is some real hard core changw, not just a rehash.

You see it's not that I don't know about the new stuff and fear it it's that I'll only adopt a change when there's a point to it.

Your dismissal of "FVWM is not a GUI window manager" makes you sound very, very biased. FVWM is a window manger (that is precisely what it is) and a WM in general is a feature of a *graphical* user interface. Simply because I'm using one old tool (FVWM) because I consider the UI to be more suited to my needs, you have dismissed the rest of what I said without bothering to read it or verify facts. That sounds very much like the zealotry I was warning the OP against.

Which means you aren't using unified cross applications tools so stuff like cut and paste for objects (what Microsoft called Object Linking and Embedding) won't work.

Firstly, that doesn't make any sense: FVWM is just a plain old window manager and takes no part positive or negative in the copy/paste mechanism. That's all marshalled by the X server (and any selection managers that happen to be listening for the right events) and is described quite well in the ICCCM. If copy/paste fails to work properly, the window manager cannot be blamed, since it takes no part in that transaction.

I do use unified cross application tools, just not the sort you're thinking of. My tools consist of thingslike make, gcc, llvm, gvim and various pipeline tools, some of which I write myself. They all work very smoothly together.

If you mean pasting objects of some sort into documents, then no, I don't have any use for those features, since my day job is not creating documents. My day job is writing code (generally on the scientific/techincal end, like machine learning, for deployment on web services or sometimes mobile devices and occasionally desktop PCs) with a side order of deep embedded development and a bit of hardware (circuits and mechanical) design.

For all of those things, I use the latest tools and techniques, so accusations of being stuck in the 1990s are clearly incorrect.

What you seem to be struggling over is that someone who is happy to ride changes and in fact does so with glee when good ones come along will not simply adopt the latest changes when they're bad.

but really, this topic is a non sequitur to the one being discussed.

No, ont at all. As someone pointed out elsewhere, GNU is more than just an OSS oranisation, they are fundementally a social justice organisation. There goal has never been "to produce the best OSS".

But these days it is mostly older folks who have a static computing environment they've been using for the past 18 years or so and want to continue using it and don't appreciate change.

Careful there, you're as guilty of the quick dismissal of other people as they are of GNOME.

I am admittedly appear on the surface to be a die-hard of the old ways. I use FVWM2. As far as I'm concerned, change is no bad thing in and of itself.

I sometimes compile the latest mainline GCC from SVN because the new versions of C++ are franlky better.

Many changes to the general DE are a big improvement: I might use FVWM, but there's a free desktop compliant tray sitting there in the mix.

xrandr and it's GUIs are great: no more editing config files and restarting X thank god.

I'm even beginning to make a major change in my life I never thought would occur: switching from the venerable xterm which has served me so well to Terminology since it's frankly an awsome new thing.

I'm also up to speed on things like the latest bluetooth 4 and even ran out and bought a BLE dev kit shortly after they became available. All good stuff.

If given the choice between a modern system and a system styled after one from when I got started with UNIX (1994) but with modern specs, the choice would be clear. The new systems are immeasuably better than the old ones in an interesting variety of ways.

Simply dismissing me as "hates change" is utterly inaccurate. I think part of the hostility GNOME gets is that if there's anything new someone doesn't like, they get instantly dismissed as "hates change". I like a lot of the new things which are present.

Sadly, in order to be relevant you have to change.

No, that's utterly untrue. You need to change some things. Blindly changing everything won't make you relevant (relavant to what even? you have to e relevant to something).

What damage are you talking about?

I'm not a gnome user here, but I can give some insight, perhaps.

There's quite a bit of general suspicion and dislike of systemd. Whether or not this has merit remains to be seen. Whether systemd is the best system remains to be seen as well. As a non gnome user, even I've felt the effects: forcing systemd to be a hard dependency of gnome means that basically all distros have had to switch because of the popularity of GNOME, not because of the relative merits of systemd. Many people consider this damaging.

I'm sort of on the fence about systemd. It seems sort of alright, but is somewhat harder to hack than the old RC scripts. It also seems rather complex and few if any of the supposed benefits have become apparant on my machines. The tying in of GNOME so thorughly does seem like a bad idea: the lack of tie in in the first place alowed a quick and almost seamless switch out of the init systems. The heavy vertical integration will make the same feat almost impossible in future.

Personally, by biggest bag of bile is aimed at the excerable file dialog box which has finally wormed its way into many programs I use. Basically it ignores the current working directory of the program bringing up the dialog box. At a stroke, it's invalidated the last 40 years of unix UX for no improvement. And by no improvement, I'm talking about literally no improvement: the experience could easily be the same as it is now for GUI only users while still working correctly for commandline users.

If you're still reading, I'll happily elaborate on this, and a couple of other things, specifically the print dialog box.

It's those two things that to me have been big usability regressions to the point where I've even vaguely pined over the old Motif dialog boxes at times.

What happened to making the best OSS software possible?

You know GNOME is part of GNU, right? Try reading a bit about the differences between OSS and Free Software.