Most used programming languages...

As a note, OS/2 is still around, still has up to date Java, and is still updated, maintained and used by many.

I haven’t followed the further development of OS/2 and its successors eComStation and ArcaOS in detail after IBM canned it (and thus made it pretty unlikely that anyone not locked into legacy stuff would continue using it), but my impression was that the last credible Java version was a community port of OpenJDK 6. The last production-ready version was IBM Java 1.4.2.

But I am sure you can provide a link where I can get the latest Java 18 Early Access version for OS/2. Not that it would run sensibly on my OS/2 Warp 486 PC I still have somewhere…

The bytecode engine in Java is not to complex

This sentence makes no sense. The JVM is the bytecode engine. The different JVM implementations range from very simple to insanely complex. I guess you meant “bytecode representation” or “bytecode specification” instead?

though it also does not lend itself to speed very well

But in practice, Hotspot and OpenJ9 are very fast, rivalling the speed of native code. And if you really want/need native code – sacrificing a bit of flexibility – the GraalVM Native Image feature covers that nicely. Not mature on all platforms yet, but gets better with every release.

and there is a tendency for programmers to create unreasonably bloated software in Java.

This tendency is there for all languages apart from those that make it very hard to write reasonably complex software. It mostly depends on the developers, not on the language used.

It is possible to create good software in Java, it is just rarely seen.

Compared to what? I very rarely see good software. Independent of implementation language.

Then there is the fact that Java is a Sun (Oracle today) product, not a universal completely open source, open spec language.

Completely untrue. Up until the release of OpenJDK and the organization of further development via the JCP, Java was not really open. Nowadays, it is. “Nowadays” as in “for the last 10 years”.

As usual these are just my personal views.

Mostly based on misinformation and/or outdated information it seems.

There is currently a new Java release every 6 months (with an LTS every 18 months). A lot of the new features are there to help developers write more readable code (ie improvements to Switch statements and function programming).

I gave up on Java decades ago because they kept on introducing ‘new’ methods, etc, faster than I could learn them! And the ones I’d learned were invariably then ‘deprecated’.

Everything I typed in this post is my view, based on ongoing daily research that will be kept up to date as long as I breath.

I am especially interested in your ongoing research wrt the claim that OS/2 has an up-to-date Java implementation available.

I gave up on Java decades ago because they kept on introducing ‘new’ methods, etc, faster than I could learn them! And the ones I’d learned were invariably then ‘deprecated’.

This is strange, because you can still compile and execute (the vast majority of) your Java 1.1 (i.e. from more than 20 years ago) code on the very latest Java version. There are very few “breaking changes” throughout the Java history – mostly accessing JDK-internal classes (hey, your IDE warned you about that!) and when new keywords are introduced. There were a few things that have been removed from the runtime, but you can usually just add them by yourself (JavaFX, JAXB, Nashorn…) to your application instead of relying on the JRE.

Of course you have to keep in mind that just because something is deprecated in Java, it will not go away. Only if it is “deprecated for removal”, it gets interesting, like it was the case for Applets, where the API will finally get removed with Java 18ff.

When it comes to backwards compatibility, only few languages care as much as Java, while still modernizing the language and the runtime at an astonishing pace. Overall, the Java universe is a pretty backward-compatible thing, and things don’t get changed without proper investigation. With the possible exception of changing the default encoding for property files of course…

So when DavidS writes…

Same reason I gave up on trying to use Java for anything worthwhile. It is still a moving target, obsoleting code faster than a rabbit cotton-tail.

…it looks like the same spreading of misinformation as ever.

When it comes to backwards compatibility, only few languages care as much as Java

Well, it’s a better attitude than PHP, and lets just not talk about the Python 2 and Python 3 debacle.

It can be deprecated AND supported.

It can be deprecated AND supported.

The point of DEPRECATED is that it IS supported – for the present – but probably won’t be after some future release, or may not be on some versions of the architecture that are otherwise compatible. (That was certainly the meaning when I was writing the docs, anyway.)

The point of DEPRECATED is that it IS supported

I’d always understood it to mean “still present and working, but if you find a bug: well that’s another reason we deprecated it, aren’t you glad the buggy bit is going?”

Sort of similar to "Google fully supports this – this week "

“still present and working, but if you find a bug: well that’s another reason we deprecated it, aren’t you glad the buggy bit is going?”

That’s very often the unwritten subtext to the official story I regurgitated, certainly.

The point of DEPRECATED is that it IS supported

Funny. My impression of “Deprecated” is the complete opposite. It is not supported, but it is still present. Hence anything that uses something deprecated won’t suddenly blow up, fail, or dump endless errors (PHP, I’m looking at you).

Generally speaking, things can be deprecated if something new and better comes along, but they should only be removed if there’s an actual security risk. If the language just decides to reorganise things or change function names in order to improve something, then there should be some sort of aliasing between the older style and the newer style. You know, like how you can still use K&R function definitions in C. Because all the existing code doesn’t rewrite itself.

And then there’s Python. Which managed to make a bigger balls up of changing stuff than .Net, and that’s saying something. Perhaps a tacit admission that the original Python was badly designed? I don’t know, but it doesn’t give me too much confidence for Python 5.

but probably won’t be after some future release

Yes, processors are permitted to have a stricter definition of deprecated (“this instruction is no longer officially supported and will be removed soon”) because there are architectural reasons why they are being removed (like SWP).

That said, some architectures toss out old unwanted stuff, others don’t. Your shiny liquid nitrogen cooled multicore x86-64 slab still boots up in 16 bit Real Mode and you can (sort of) boot up MS-DOS and the like if you wanted. Native, not under any emulation. The problem is that such old systems don’t understand modern hardware (EFI boot, USB keyboards, SATA, NTFS on ridiculously huge drives, etc etc) and if there’s no BIOS compatibility built it (Intel are starting to remove this) then you would need to develop your own drivers and good luck addressing stuff in a 16 bit world. ;-)

Sort of similar to "Google fully supports this – this week "

Given Google’s past history, that can count for pretty much anything that isn’t Search, YouTube, GMail, or ad-flinging. https://killedbygoogle.com/

Yes, processors are permitted to have a stricter definition of deprecated

Oh, I spoke only of how ARM defined DEPRECATED up to at least 2007 – and only in the context of the instruction set architecture at that. I don’t recall using it in any other context. Whether other organizations defined it the same way, or at all, I couldn’t say. I wouldn’t even be confident ARM still define it exactly that way; things at ARM have changed a great deal since I left – not, I hasten to add, connected in any way with my retiremnt!

they should only be removed if there’s an actual security risk

Oh, perhaps in software that’s fair enough. In hardware & computer architecture, there are other valid reasons for removing something – things I was conscious of were reusing instruction set space (think the Never condition in early ARM architectures, but there are other (smaller) examples), power saving, chip real estate saving, and streamlining instruction decode. Keeping the odd bit of old baggage is one thing, but burdening yourself with loads of it (Intel style) is another…

but burdening yourself with loads of it (Intel style) is another…

Could explain why the ARM is itty bitty and x86-64 devices are, well… Those heatsinks look like something out of Starfighter 3000.
https://c7.alamy.com/comp/R46RAM/new-processor-ryzen-7-2700x-and-computer-motherboard-asus-rog-crosshair-vii-hero-dd4-and-nvidia-gtx1080-R46RAM.jpg

Looks like the same attempting to discredit the truth as ever for you.

Your posts would be a lot more credible if you’d actually start to substantiate your various extraordinary claims you are happy to liberally throw around. Maybe others could then start to see where your own personal “truth” comes from.

The specification did (and to a more limited extent still does) deprecate many things, the fact that the implementation still supports much of what is stated to be deprecated is besides the point at hand.

OK, now please look up what “deprecated” actually means in Java (JDK, JRE, i.e. the SE platform – obviously not in any old library that might break compatibility any time) world instead of fantasizing about what you THINK the meaning is. Hint: it is completely different to “deprecated forRemoval”.

Even better would be if you come up with an actual example where Java changed in a way that it “obsoleted your code”. Going by your comment, there must be a whole lot of examples for that, so that should be very easy. Oh, and don’t forget to answer the OS/2 question.

Using a term should mean what the term means.

That would be all very well if terms had fixed definitions, but English isn’t generally like that. Thus when complete clarity is required you have to define the terms you use. ARM used to be (and may still be, I’ve not checked) very good about that.

“Unsupported”, like “Deprecated”, does not have a very precise meaning in English. You need to check exactly what it means in the document you read it in.

I don’t recall using “unsupported” in the ARM ARM, or in the ARM Assembler manuals, so I suspect it’s not a term ARM uses (or used in those days), but it’s possible it’s just a hole in my memory. If we did use it, there’ll be a definition of it in the books.

I’ll cite the definition in the ARMv8 instructions set document in the glossary =>

Deprecated – Something that is present in the Arm architecture for backwards compatibility. Whenever possible software must avoid using deprecated features. Features that are deprecated but are not optional are present in current implementations of the Arm architecture, but might not be present, or might be deprecated and OPTIONAL, in future versions of the Arm architecture.

You can even find the definition of OPTIONAL there which makes it a bit more complicate or clear ;-) I guess that definition might work also for programming languages…

But within the document itself sometimes it says just deprecated or also deprecated due to performance reasons…so still a bit unclear and really seems to depend on context…

Optional in the ARM definition is due to features of the ARM chip being optional to implement by various manufacturers. A chip can be made with or without NEON for example.

Deprecated simply means ‘will be made obsolete without further warning but not yet’.

If you program in a language, you should not use deprecated things. The next time you update your compiler, it may not compile because they have moved from deprecated to obsolete.

If you write a compiler or interpreter, you must support deprecated items.

This way, there is an overlap. A warning for programmers that some things are going to go away whilst they are still working.

I’ll cite the definition in the ARMv8 instructions set document in the glossary =>

I think that’s a more recent definition than 2007, isn’t it?* I don’t recall us actually defining OPTIONAL back in my day…although Neon was optional from its beginning, so perhaps I’m just misremembering.

“When I use a word,” Humpty Dumpty said, in rather a scornful tone, “it means just what I choose it to mean—neither more nor less.”

Absolutely so. Lewis Carroll understood the nature of the English language extremely well, and was a real nerd of his time.

* Edit: well, it obviously is, since it’s the ARMv8 document. But I think it’s also a bit more detailed than the 2007 ARM ARM’s was.

Rick wrote (while I was in Cornwall with no internet):

And then there’s Python. Which managed to make a bigger balls up of changing stuff than .Net, and that’s saying something. Perhaps a tacit admission that the original Python was badly designed? I don’t know, but it doesn’t give me too much confidence for Python 5.

Yes it was an admission that Python 2 was badly designed, particularly in the handling of Unicode. So they had a choice between letting the language slowly die through creping irrelevance, or take the pain of making some incompatible changes to fix it with Python 3. As there is no Python 4 yet, it’s a bit too soon to make assumptions about Python 5.

You could say exactly the same thing about ARM and the 26 bit to 32 bit mode breaking changes. If they had stuck with a 26 bit processors restricted to 64MB of program memory, they would be no more than a niche low power micro controller supplier today.

Python 4 will probably be backwards compatible with 3 to a greater degree than 2, but it may not. If ARM hadn’t developed the non backwards compatible 64 bit ARMv8 ISA, they wouldn’t be supplanting Intel in data centres, and would probably only be in the budget end of mobile phones now.

With both Python and ARM, there comes point where absolute backwards compatibility becomes less important than changing to meet new challenges, while maintaining the same design philosophy and tooling. The alternative is to be replaced by something completely different, which requires even more relearning and porting.

in part the opinion of the poster.

Only in part?

The quality of a new OS can be seen from

I’m going to disagree. Surprise! 😋

In the late ‘70s, you needed a degree (probably in electronics) in order to use a small “home” computer. They usually came in kit form, had to be assembled (and debugged), and some of the friendlier ones provided a monitor that allowed you to enter a program by bashing in hex codes. We won’t even talk about the price. Accordingly, they weren’t very popular.

In the eighties, home computers got a lot friendlier. A little box that plugged into a TV, cassette tapes, and some form of (usually pretty lousy) BASIC. These machines were often quite simple both electronically and in terms of firmware, so one person could understand the operation of the entire machine. Prices were a lot less scary, and in the UK the push via The Computer Literacy Project helped introduce a lot of children to computers. They were very popular as they had potential, and it was a pretty even playing field – a smart teenager could put together a game every bit as good as commercial releases.
The infamous ZX Spectrum was a piece of shit for so many reasons, but it was instrumental in getting a lot of kids (who had plenty of free time) into computers and programming and such.

In the late 80s and 90s, machines sort of lost their way a little. More capable, more powerful, but too often viewed as “things for games”. At the time Amstrad made some machines that were less a computer and more a “business machine”, it would start up with a world processor or spreadsheet… They weren’t terribly good, but Amstrad did figure out that what was putting people off was that oftentimes computers were “too complicated”.

I managed to program my mother into understanding that A: was the little slot on the front, C: was something inside, and D: was the CD-ROM.
Then she used my PC where D: was the second harddisc and was lost…

It wasn’t until the mid 90s (with the introduction of Windows 3.1) that computers started to get more popular. It’s pretty easy to malign Windows for a billion technical reasons, but psychologically they made things a lot simpler. The machine could be set to start up Windows automatically, so every subsequent interaction could be pushing the arrow to little pictures and prodding a button to activate that picture. Applications such as Word that, at the time, weren’t that complex but coupled with a laser printer (increasingly less expensive) could output professional looking documents that were a million times better looking than the dot matrix efforts of the previous decade.
Better yet, Windows was pretty much device agnostic. You didn’t need a “Mac” or an “Acorn” or an “Amiga”, you just needed something “PC compatible” within your budget. Windows requirements, in those days, was really modest. So much so that it can run slowly but acceptably under DOSbox on a Pi!

Therefore, it is extremely clear that the popularity of a machine (and its operating system) has nothing to do with how many people learn to program it. Indeed, when considering the world at large, most people don’t give a flying figlet about programming. They really don’t. Their entire justification, from about the days of DOS 3.3 onwards, is “can this thing do what I want?”.
The more computers were able to answer this question, the more popular computers became.

In the late 70s, maybe one computer per town.
In the early 80s, the cool kids had a machine. Smug Beeb owners, or the riff-raff with their Speccies.

Fast forward to today, an average modern household may have at least a laptop, maybe a tablet or two, and pretty much everybody has a smartphone (y’all can argue amongst yourselves Android or iOS, because everything else got left behind).

In terms of commerce, a successful machine is defined simply as “one that sells”. Cheap to make, easy to sell, makes profit.
It’s perhaps worth noting that in terms of volume and user base, every single example you list added together cannot hold a candle to a generic mid range smartphone.

How many Beebs were sold? Or the AppleII? Or the Atari?

Now pick a smartphone manufacturer and a reasonably decent (but not flagship) model.
When computers can do things that people want, they’ll buy them. It’s as simple as that.

I rest my case.

many too young to have a degree in anything.

Here’s a hint – when you’re talking with Brits, expect some degree of sarcasm. It’s just how we are wired up.

Who said anything about popularity?

It’s how I interpreted your talking about how “well received” a system is. What is well received of not a synonym for popular?

Plus – while there are some exceptions – operating systems, computers, even entire companies lived and died depending on their “popularity”. The world’s best system that nobody uses is indistinguishable from a polished turd that nobody uses. What they have in common is that they both, essentially, failed.

And the definition of success is and always has been subjective.

True, but I think the commonly accepted definition is “coin comes, not goes”. You might not subscribe to the commerce definition, however a computer that is made in order to be sold is pretty much exactly what commerce is so no other definition makes sense. A company making profit is generally considered good if that allows for development of newer better things.
[though a company making profit to drop the money into the pockets of shareholders is rather less good, that’s the point when it stops innovating and becomes little more then a gravy train]

even have crap outsell much better products, that are more affordable, in the same area.

ßetamax vs VHS, just to demonstrate “it’s the same old story” (everywhere I go, I get slandered, libelled, I hear words I never heard in the Bible….).

ßetamax vs VHS, just to demonstrate

If you add in the V2000 standard then I think you get a nice progression of quality going one way and availability of content, and hence sales quantities going the other.

Do you remember the Pet Rock? A good example of something extremely popular only because someone was good at marketing it, it was completely useless otherwise.

Something about that made me think of Trump. :)

:-)