On Sun, Aug 25, 2019 at 04:56:28PM +0900, YW LEE wrote:
> >
> > I think, it would be much much better and it would be much appreciated, if
> > I could hear directly from you what you wrote here as above, in your reply
> > email to my cordial invitation email, instead of you being silent or you
> > saying something else in your reply.

I did tell you all of this (perhaps in a bit more polite fashion, in
terms of my opinion of the ISO standards body).  But if you like, I
can dig up my messages from my mail archives.

I explained several times, both in person when we had the chance to
meet in Korea, as well as email.  I distinctly remember explaining
current funders of the Linux Foundation didn't see the business value
in continuing to fund LSB work, and explained how the technology far
too fast moving for ISO to add value, which is why things like Docker
images are not being standardized in ISO, and you refused to take the
point.  I eventually simply gave up.

> > < I tried asking Professor Lee for what he thought the business case and
> > < the market for the LSB, and he couldn't give me one.
> >
> > As far as my memory serves me right, I vividly remember that I explained
> > much and very clearly for the business case and the market both in the face
> > to face meeting and in my emails to you.
> >
> > There, I clearly and repeatedly mentioned, Smart city, Smart home and
> > Smart Society, Smart Car and many other things which are related to the 4
> > th Industrial Revolution, Internet of Things, Cloud Computing, etc as the
> > business case and the market for the LSB.

All of the above are buzzwords.  Merely including a list of marketing
buzzwords isn't a business plan.  *How* does standardizing an Linux
ABI help for making things easier for "Smart home" or a "Smart Car" or
"cloud"?  OK, a company is interested in say "IOT" or "Smart Cities";
how does investing resources in LSB standardization provide a positive
Return On Investment for that company?  A lot more specificty is
required.

"Smart Home", "Smart Car", and IOT are all embedded use cases.  In
those cases, the software is embedded in the product which is sold to
the end user.  In the case of open source software, the ABI doesn't
matter, since the source code is provided from the upstream provider
to the product manufacturer.  Examples of this include the Board
Support Package provided from Qualcomm to support its ARM chips, and
the Android Open Source Project source code from Google to the various
Smart Cars, Smart Home, etc. devices using AOSP.  ISO standardization
provides *zero* business value for these cases.  If you think it can,
you need to exaplain exactly what ISO standardization enables that
can't be done simply by shipping open source code.

If there is proprietary code involved, say, such as for example,
Waymo's self-driving car technology, that can be provided as source
code under NDA to various car manufacturers.  Between copyright law
and patent law to provide intellectual property protection, there is
no problem providing proprietary source code to a small number of
product manufacturers.  Again, it's not enough to list "Smart Car" as
some magic phrase.  That's not a business plan.  What does ISO
standardization enable that can't be done by simply shipping source
code around?
 
The LSB made sense when the goal was to enable proprietary software
vendors to be able to ship software in binary form to end users.  If
you need to ship proprietary software to thousands or millions of end
users, then shipping source code around in NDA doesn't work, since the
chances that the software will leak is too high.  The original hope
was that the LSB would enable software such as Intuit's TurboTax to be
made available to enable the Linux desktop case.  The idea was to
enable a software vendor to be able to ship their binary software
independent of what distribution might have been chosen by the end
user.  The business case, then was that by enablng a richer ecosystem
of Linux ISV's, it would allow companies like IBM, HP, SuSE, etc., to
be able to make more money by growing the overall market, and hence
increase the size of the pie that could then be split amongst those
companies.  This is a lot more detail than just saying "enterprise
software" --- we had an explicit idea of how we could enable companies
could find it in their own self-interested way to support the LSB.

Unfortunately, we couldn't succeed quickly enough, and in the mean
time, Red Hat took a commanding lead in the enterprise software space,
such that Red Hat judged it not in their interest to cooperate by
growing the overall pie, but instead found it more economically
rational to try to completely grab the entire pie.  And the desktop
market for Linux never materialized.  Since then, web applications are
not much more of a thing than they were 15 years ago, and the rise of
phones are such that most people have an iPhone or an Android devices,
and so many of the desktop applications have now become mobile
applications --- indeed, many users don't even own computers, and only
rely on their cell phones.

As far as "Cloud Computing" is concerned, a lot of software that is
run on Cloud computing is open source software, or company-specific
business logic.  And for business applications, who cares about Linux
standardization when companies can simply ship their software bundled
in a VM image for AWS or GCE?  

> > < He pointed me at some ISO web site that talked about how standards were
> > good
> > <    for Toilets and the Plumbing industry.
> >
> > Again, as far as my memory serves me right, I remember that I never
> > mentioned like what you wrote above. Not true! Not a fact at all! Your
> > mentioning toilets as above could be a kind of humiliation to me. But,
> > rather, I would like to believe that you are full of humor, thus you wrote
> > like above.

When I tried explaining to you that I saw no business value in
standards in the computing business, you pointed me at this web site:

https://www.iso.org/benefits-of-standards.html

It was a generic web site about how ISO provides business value, and
wasn't specific to the computing industry at all, let alone the Linux
ecosystem.  It was about at this point, that I decided it wasn't worth
trying to argue with you about why I didn't think there was any
business value in standardization, at least not in the fast-moving
open source and computing world.

> > Mr. Linus Torvalds himself kindly recommended and introduced you to me
> > there. There, Mr. Linus Torvalds told me that you are a faithful and
> > reliable expert regarding LSB and you were a key member of the last LF’s
> > LSB project which produced the existing ISO/IEC LSB standard, the ISO/IEC
> > 23360: 2006(release year) series, and, at that time, you were at IBM, which
> > provided financial support to the LF’s LSB project to make the ISO/IEC
> > standard. So far, I have paid you my best courtesy and my sincerity. I
> > thought that you loved LF’s LSB or you were a big fan of LF’s LSB.

IBM had very specific business goals in supporting the LSB two decades
ago, which I alluded to earlier in this e-mail thread.  This including
trying to make sure there was a counterweight to Red Hat for the
enterprise space, and because selling to certain European government
entities would have been easier to do if IBM sales people could point
to an ISO standard.  But that's history.  Since then, Sun Microsystems
has gone out of business, and most of the Legacy Unix OS's have been
abandoned, and even Microsoft has acknowledged that Azure has more
Linux OS's running in their cloud than Microsoft Windows.  The need
for the ISO label as marketing has largely evaporated.

I very much cared about the LSB; as I said, I was around during its
earliest days, and was on the initial board of the Free Standards
Group both when it was initially created, and when the FSG merged with
the OSDL to form the Linux Foundation.  But that was because I wanted
Linux to be successful, and I was afraid that Red Hat was throwing its
weight around too much, and dominating the ecosystem.  I wanted other
distributions to be able to have an even chance.  But that was then;
this is now.  The reality is that Red Hat has ended up dominating the
enterprise space.  Our goal of trying to avoid that outcome via the
LSB was a failure.  However, the enterprise space is no longer the
major Linux market that it was then, and the shipping of Linux
executables that will work on multiple distributions is not as
important in 2019.  In fact, one could argue it's barely important at
all.

There was some very good work that was created by the technical teams
who worked on the LSB, both at various companies who contributed
engineers to the LSB working group, and our contractors at ISP RAS.  I
was very proud of the technical work behind the LSB.  The failure was
not in the technical effort, but in the business value returned by
that work.  And whatever reasons may have existed for companies to
support the LSB in early 2000's, they are even weaker or non-existent
in 2019.

Which is a shame.  I would have liked to have been able to buy copies
of TurboTax and Adobe Photoshop/Lightroom for Linux.  But these days,
Adobe is trying to move its Lightroom users to a model where you use
Android and IOS applications, and where the actual image manupulation
is done "in the cloud", in Adobe's data centers.  In that cloud
computing model, the LSB is irrelevant for that use case.  Times
change, and how applications are delivered to end users have changed
as well.

Best regards,

                                        - Ted
_______________________________________________
lsb-discuss mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/lsb-discuss

Reply via email to