Continuum Analytics upsets me with Anaconda 2.0

121 views
Skip to first unread message

anatoly techtonik

unread,
Sep 6, 2014, 11:25:41 PM9/6/14
to spyd...@googlegroups.com
...PyQt is now the default Qt binding, as some users were experiencing stability problems
with spyder...
http://continuum.io/blog/anaconda-2-released

I feel like claims that Anaconda is completely free now are false

Why am I upset about it? With so much buzz and marketing power that Continuum
delivers, I expected guys to be active somehow in the open source part of the ecosystem
and pay attention at the values of the project that are not only at the top of their IDE list
but also important enough to break dozens (hundreds?) internal application. That's why a
major version change.

I am not saying that everybody here wants non-restricted Qt bindings on Windows. People
usually don't care, but after MS killed Nokia, so that there is no PySide team anymore,
I'd expect a company like Continuum to be able to calculate the impact made and provide
at least some support value back to PySide project. Reporting bugs at minimum.

It is just my personal rant, but...
Getting the best out of open source projects, wrapping them in package and marketing it
at a conferences. This is not what you expect from a scientific company that holds the
keys to open source, algorithms and processing. You expect them to be on a edge of
researching the economy, the system that powers it - the ecosystem - to make sure that
useful agents survive, not die. You expect them to be leaders that explain the trends, how
the stuff works, to provide some hope for this darkness. And what you see from the
ecosystem point of view? Just another parasite trying to survive in this "economy".

Just to make clean about the matter. I am sitting right now trying to code some stuff for
the money that will be plenty enough for food or buying new clothes, I see people quitting
social science, biology, neural networks labs just to earn cash and write dumb 
games or join outsourcing business (because, well, it is boring for US developers to write
and maintain code for their own products). My "quality of life" directly depends on the
amount of people involved in research jobs, because I hope that one day it will be possible
to find a solution for my personal issue. Hoped. If money always kills the game, it is
pointless for me to continue in the open source race, because I will never be able to
afford the costs of the outcome as I am not as smart as others to earn some.

Ronan Paixão

unread,
Sep 7, 2014, 4:44:03 PM9/7/14
to spyd...@googlegroups.com
I don't understand why you're so upset. First of all, Anaconda 2.0 was not released now, but in may. Then, what I found out is that Spyder was one of the applications that weren't playing nicely with PySide. Migrating did solve most of my problems by migrating to PyQt. Besides, it just changed the default Qt package and PySide is still available through conda.

Also, I don't think that Continuum is "just another parasite". It's making Blaze (though on the shoulders of NumPy, but with parallel in mind), conda and Numba. Although the Pro versions are paid, just releasing some of that as open source seems to me like "non-parasite".


--
You received this message because you are subscribed to the Google Groups "spyder" group.
To unsubscribe from this group and stop receiving emails from it, send an email to spyderlib+...@googlegroups.com.
To post to this group, send email to spyd...@googlegroups.com.
Visit this group at http://groups.google.com/group/spyderlib.
For more options, visit https://groups.google.com/d/optout.

anatoly techtonik

unread,
Sep 8, 2014, 1:50:56 AM9/8/14
to spyd...@googlegroups.com
Migrating to PyQt solves problems only for GPLed products, otherwise
you need to pay license fees for non-GPLed version of the binding on
Windows. Continuum is a parasite in Spyder ecosystem, because they
deliver stuff made by others without contributing back. Although they
donate to NumFocus, which lists Spyder among list of its projects, I
don't really know how NumFocus really helps Spyder. So despite all
marketing buzz, money don't work here. If there are bugs in PySide,
they should be at least reported so that people had a chance to fix
them. Continuum could raise awareness that open source projects that
it delivers also need bugfixes and contributions. But they don't do
this in their marketing materials. They use and promote, get attention
and resources, but then they prefer to hide the issues. Why hide?
Because people perceive that Continuum plays a major role in the
ecosystem and if they can not handle the issues, the image will be
hurt and market share will fade. That's just an assumption. I am not
saying that they should pay everybody, but they the only company that
can choose how they deal with things, and who pretends to do analytics
well, so I believe that they definitely could find an alternative way
to settle things in a way that is beneficial for all agents.

As a consequence of Continuum action, therefore visibility of the
actual problems with PySide becomes even lower. There are little
chances that issues in PySide are be fixed if companies that use them
continue to ignore the fact that project need maintainers and skilled
people to tackle stuff on C++ level. They prefer that "somebody else"
will fix issues automagically. And it happens. If people feel that
this is important. But Continuum doesn't even do this. What message
this Anaconda 2.0 release conveys? "Spyder is cool, we are delivering
it, but PySide is a pain in the ass, so we turn it off". PySide is an
awesome project, but it requires some support too, and who if not
commercial vendors (who own developers) should think about how to
solve this problem, then who? The message could be "damn, the PySide
brings us down, there are multiple issues with .., .. and .., which
can be improved like .., ..". This is what I expect from guys, who are
not parasites.

I am not a user of conda, Numba and Blaze, but I am the one who picked
up Spyder in early days and learned Qt with its code, so its code is
already in my head, and even though I am not active now it is still
damn sad to see that people are unable to do anything bigger than just
"use" the software instead of trying to improve it for everybody. You
know, when you're contributing to open source that is done under
non-restrictive license, you do so because you value the effort that
was made so far by these people to bring this stuff live for all us to
use, and you just try to pay back improving it a little for everybody.
But now people are just taking the existence of such things for
granted, or they just don't feel capable, or corporate environment
denies their "resources" from "wasting" time on doing this. In any
case this a sign of regress. People are reinventing bicycles instead
of trying to improve existing things and this is a waste of time on a
global scale.

Maybe my expectations for scientific company are too high, but if not
them, who else can do anything? Big business is not interested,
because they need to see money, not to "increase waste" to feed 85% of
their loyal workers who are "not providing any significant value". I
hoped that Continuum is different and tries at least explore what is
this new "knowledge economy" buzzword is about. I've heard that Europe
promoted it as our future. But if awesome open source projects like
PySide are dying, I don't see any hope for open science or anything
open at all.
> You received this message because you are subscribed to a topic in the
> Google Groups "spyder" group.
> To unsubscribe from this topic, visit
> https://groups.google.com/d/topic/spyderlib/gZLrhwUGRHk/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to
> spyderlib+...@googlegroups.com.
> To post to this group, send email to spyd...@googlegroups.com.
> Visit this group at http://groups.google.com/group/spyderlib.
> For more options, visit https://groups.google.com/d/optout.



--
anatoly t.
Reply all
Reply to author
Forward
0 new messages