[Interest] [Development] Windows 7 support will be dropped in Qt 6

Sun Jun 14 19:55:20 CEST 2020

On 6/14/20 5:00 AM, Jonathan Purol wrote:
> One possible solution that came to mind is that after the LTS is dropped
> for Qt 5.15 it could be opened to the public for contributions with some
> elected people as the maintainers, perhaps from larger companies that
> are in close contact with the Qt company.
> That way there would still be support for Qt 5.15, basically
> indefinitely, without every single company needing to invest the same
> amount of effort just to have a maintained version of their software
> running.
> That said, I am unsure if this would be a choice the Qt company would,
> and even could make. There might be licensing issues or other things
> prohibiting it.
> Additionally there could be problems with the direction of the project.
> Version 5.15 being controlled by few big players might lead to some very
> off-scope specific features being implemented which could cause hassles
> for less influential groups of people.

Your suggested compromise has already happened.

Qt was forked some time back as CopperSpice. https://www.copperspice.com/

It has diverged quite a bit since then. I'm only just starting to play 
with it as their licensing is "much goodlier."

Requires a C++17 compiler. Got rid of MOC. They don't physically state 
it, but the documentation has no mention of QML or JavaScript and 
dropping both of those would be a very good thing.

Supported platforms appears to go all the way back to Vista for Windows.

https://www.copperspice.com/docs/cs_overview/supported-platforms.html

I don't see Centos 5 on here which was mentioned in this thread, but 
given the wide range of gcc versions I'm guessing nobody tried until 
Centos 8 and you probably could get things to compile for Centos 5.

After I get done tweaking the Diamond editor to support themes and EDT 
keypad mode I want to try the Migration information out and port my 
XpnsQt project to XpnsCS.

https://github.com/copperspice/diamond

https://www.copperspice.com/docs/cs_overview/migration.html

https://sourceforge.net/projects/xpnsqt/

I wouldn't go so far as to say I expect the learning curve to be quick, 
especially since I'm helping with the farming now. It shouldn't be like 
starting over with a new library though. Just a lot of the really bad 
parts of Qt were removed and some new features added. There is a series 
on my geek blog for those interested.

https://www.logikalsolutions.com/wordpress/uncategorized/copperspice-and-diamond/

There can't be compromise Jonathan. It's a binary situation.

You either fully support Windows 7 for the embedded/industrial crowd 
that will be using it for at least another decade, or you don't and they 
stop paying, contributing, hiring, etc. when it comes to Qt.

Or you don't fully support it. Only half-ass it, kind of like happened 
with C++11 being supported as "the minimum" then when some fool (me) 
tried to compile using that standard on three different major Ubuntu 
releases he found that it didn't work and hadn't worked for quite some 
time. People using things that only existed in later standards version 
and they were just being checked in after code review because nobody 
ever built using C++11.

If one tries to "compromise" Windows 7 will end up going the exact same 
way in under six months and like far too many bugs that don't have to do 
with new and sexy, the bug will be allowed to rot in the backlog.

No development library can be all things to all people. Qt has been 
trying to do that. This is where it failed. It can either target "the 
cool kids toys" meaning they will have to do massive redevelopment every 
two years and bugs won't get fix -- OR -- they support "real industry" 
where products have lifespans of many decades requiring some support 
staff for their entire life which means if you have a reasonably priced 
support contract, you get many decades of revenue provided you actually 
do support and focus on the bug backlog.

(It's about time for "that guy again" to pop up. The one paying for a 
support contract who has critical bugs that have been in the backlog for 
years.)

I haven't physically spoken with anyone at CopperSpice, only done my own 
extremely limited research (and that blog post series will show just how 
limited) but the thing that intrigued me most of all was their 
"Subscription" pricing.

https://www.copperspice.com/subscriptions.html

For me personally, the bronze level of $100/month per seat is too much. 
For me as a contractor working on a medical device for someone, 
$100/month is more than reasonable. Some of my clients will probably 
bump to the Silver subscription of $350/month. I don't walk in the 
circles that would pay the $800/month 20 seat minimum subscription for a 
year, let alone years.

We always see "Just use 5.15." What I and others have been trying to 
convey here is that statement and many like it are made by people who 
don't understand real business. Real business isn't Apple.

I lived through this with the transition from Qt3 to Qt4 and the 
sweeping API changes followed by dropping of OS/2. "Oh, just use Qt3." 
Well, that medical device is still being made, still saving lives, and 
contracts for Qt3 with OS/2 seem to hit my inbox every 12-20 months. The 
kids keep making sweeping API changes so you can't even take a new Qt5 
developer and have them get up to speed in any human amount of time with 
Qt3.

I can take someone who just graduated college learning the latest and 
greatest COBOL standard, teach them the 72-column "card" format of 1970s 
standards COBOL and have them up to speed in an afternoon. "Real 
business" can't do that with Qt because it has spent too much time 
trying to be all things to everyone instead of minding the store.

Now, Qt project and Qt Company have to make a decision which markets 
they are going to kiss off. I can't even believe Scott and his company 
are both still here and still paying money to Qt Company after the 
number of times they've been screwed by both.

If they choose to dump Win 7, it won't be just those products. The 
companies getting screwed by that decision will take __all__ of their 
contracts and money for all other platforms they are supporting and give 
it to someone that actually appreciates them.

I personally don't have any Windows 7 stuff to support. I simply have 
watched companies get screwed over and over and over again by kids 
wanting to play with the cool toys. I don't know just how much WinCE/Qt 
stuff CAT has in the field. I know they can't stop supporting that 
equipment until the mine it currently is in gets played out. Then, if 
and only if the equipment doesn't get moved to a new mine, they can 
stop. You simply cannot tell someone who spent north of $750,000 on a 
single piece of equipment the reason you can't fix it is the kids wanted 
to play with the new iPhone.

No, I'm not just being a cranky old man, I'm explaining how real 
business works.

The project and Qt Company has to make a choice now. They either chase 
the phones and kiss everyone else off, or they kiss off the phones. 
Automotive is going to have to go where the rest of the embedded systems 
goes because, unless the law changed, they have to support stuff for 20 
years as well.

Ah! The law has changed.

https://www.justanswer.com/consumer-protection-law/66qyn-does-car-manufacturer-obligation-provide-parts.html

"The dealership is not required to stock the parts but the manufacturer 
is obligated to have parts for 10 years after they discontinue a model, 
and then after that if it was not a popular model they are not required."

Not the model year, the model. So in another 10 years Ford can stop 
making parts for every Taurus they ever made.

-- 
Roland Hughes, President
Logikal Solutions
(630)-205-1593

http://www.theminimumyouneedtoknow.com
http://www.infiniteexposure.net
http://www.johnsmith-book.com
http://www.logikalblog.com
http://www.interestingauthors.com/blog