On 10/2/21 7:47 PM, Ulf Hermann wrote:
I might be wrong with those steps because I don't know the approval
process. Yet, I'm sure there is some pragmatic way to produce what you
want. You may want to share your ideas on what it actually takes.
There is not a pragmatic way within what Qt has. To start with one
cannot use Git or AGILE.There has to be a full formal SDLC following the
approved format prior to starting the coding of each release. There has
to be a coding standard heavily mirroring the BARR standard. It's a
highly controlled and regulated world and for good reason.
https://barrgroup.com/embedded-systems/books/embedded-c-coding-standard
and you cannot start with a code base that has bugs over a decade old.
While all of this is possible, it obviously is a lot of work. If you
want to do the work yourself, let's discuss the details here.
As to how to *remove* the bugs the process is simple and from what I
hear the same thing being written into the formal FDA regulation
proposal. Time consuming and resource intensive, but not complex. I will
state it one more time, at least my understanding of it, but the
regulation will formulate it into something as good as law.
1. identify bug with semi-repeatable use-test-case.
2. test with previous "official" releases backwards in time until it is
no longer semi-repeatable.
3. walk forward one check-in at a time from whatever was included in
the formal release.
4. when you hit the check-in that is either the straw that broke the
camel's back or introduces the shiny new bug, remove it from the
code base.
The process doesn't care about what came after or how many license sales
were based on that particular check-in. The process only cares about bug
amputation.
This is the process followed by pretty much every regulated industry.
Generally removal also involves some number of people being sent into an
isolated universe in some dark little room to try and salvage some
portion of the code which now must be amputated, perhaps even fixing the
bug, but regulation focuses on amputation. The FDA and other regulated
environments choose removal of a sales pitch item over the lives of
hundreds/thousands/unknowns each and every time.
As far as "cleaning up Qt" that is already being done by quite a few
organizations that have left Qt completely due to the behavior of Qtc.
See below.
If you
want to pay for such work to be done, you may want to get in contact
with the Qt Company.
Please. Qtc has been coming onto the OpenSource mailing list for years
beating that same drum. It has a high pitched tinny sound that is
harmful to the ears of human and animal alike. Removal of Qt 5
OpenSource LTS was the final straw for many. *All* OpenSource projects
of any significance have OpenSource LTS versions. Removal of the LTS
told businesses Qt is no longer an OpenSource project of any
significance thusly should not be used and businesses should not allow
their employees to contribute to it on company time.
The sad thing is Qtc is trapped in a 1980s business model. They should
really take a good look at Eiffel to see how that model works out in the
2000's.
https://www.eiffel.com/
https://www.zoominfo.com/c/eiffel-software-inc/139314020
Do not even have a D&B listing
https://www.dnb.com/business-directory/top-results.html?term=Eiffel%20Software&page=1
Kind of sad given the lone wolf operation Eiffel Tower Software does
https://www.dnb.com/business-directory/company-profiles.eiffel_tower_software.7936880af69f1c39f36fad01081c8d0d.html
It's previous iteration had a "modelled" sales figure of $3.56 million
https://www.dnb.com/business-directory/company-profiles.interactive_software_engineering_inc.ed26de719e4b5a2e0c1bd97489dbfa67.html
It doesn't matter how good you think you are doing right now, that's
where such a business model ends up.
Today's companies don't pay royalties. That business model died with
products like .RTLink and is exactly why those products died. Today's
companies use OpenSource. Sometimes they donate money. Other times they
allow their employees to contribute to the project during working hours.
Sometimes they actually purchase reasonably priced support contract.s
The days of being able to sell tools that are mostly just libraries and
live a lifestyle one would like to become accustomed to are long since gone.
Most of the "commercial" operations behind OpenSource projects make the
bulk of their money selling project level software development to
businesses.
*The Digia iteration of Qtc sank that boat for all time.*
I was brought in to sweep up behind a Digia debacle that a client paid a
ton of money for. They were paying $250/hr for my services 100% remote
and considering it a deal compared to what they paid for unusable trash
previously. Because a state machine class had just been added to Qt the
"consultant" from Digia tried to use it (most likely to put a "win" on
the Web site) for a problem that was most distinctly __not__ a state
machine. A tiny subset of the problem could be forced into a state
machine, but to do the rest of the project one had to start over.
This story from that time was not unique. Businesses paid a whole lot of
money to get a whole lot of nothing. One of the people Digia screwed in
that deal was someone Google/alphabet pays to spin things up for them
over and over. Had you made right on the deal the financial floodgates
would have opened and Microsoft would have tried to buy the company. As
it stands now the "pay Digia money to develop software" business model
is permanently destroyed. The biggest of the big players aren't going to
let a company name change hide the history and they aren't going to let
a startup they fund use the product.
Conversely, tool vendors that create entire ecosystems around a compiler
and language while providing high end high quality consulting services
do really well.
https://www.synergex.com/
Anyone could write a DIBOL front end for Gnu or LLVM but they don't.
There are OpenSource DIBOL tools.
https://www.synergex.com/open-source/
Being a private corporation the D&B numbers listed here seem low to me.
https://www.dnb.com/business-directory/company-profiles.synergex_international_corp.8c97bb50901f678b14b3590aafedaa75.html
Synergex software is used to create the banking software running the
bulk of all credit unions in America. That is a multi-billion dollar
industry. They have a reputation for developing high quality
soup-to-nuts software solutions. The zoominfo appears more accurate.
https://www.zoominfo.com/c/synergex-international-corporation/102580840
Just my 0.0002 cents
_______________________________________________
Interest mailing list
Interest@qt-project.org
https://lists.qt-project.org/listinfo/interest
