[Beowulf] HP Moonshot also with AMD Kyoto

Joe Landman landman at scalableinformatics.com
Sun Apr 14 12:07:57 PDT 2013

This is getting very non-HPC/Beowulf-ish.  My last public post on 
this.   Email me offline if you want to continue the discussion.

On 04/14/2013 01:53 PM, Ellis H. Wilson III wrote:
> On 04/14/2013 12:23 PM, Joe Landman wrote:
>> On 04/14/2013 10:31 AM, Ellis H. Wilson III wrote:
>>> I think this is strongly dependent upon whether or not the audience of
>>> the list is already exposed to the media outlet the article is from.
>> Hmmm ... a natural extension of this argument, not requiring any reach
>> at all, is that posting other published digital copyrighted work online
>> in any forum that may not have exposure to that work is ok.  Literally,
>> all we have to do is substitute "published digital copyrighted work" for
>> the word "article" and we arrive at the same point.
>> And that argument has been fairly well litigated, and the results are
>> fairly well known.
> You will note that I separated my points into three separate veins:
> 1. Conjecturing on the impact of reposting full articles in an email
> list on the aggregate traffic for a site
> 2. The legal implications doing so are likely to have, and how snippets
> deviate from full re-posts
> 3. How all of this fits into the "etiquette" and policies of a list.
> I /never/ correlated my argument that increasing traffic with making
> doing so "ok" from a legal standpoint as you suggest, or at the very
> least, that was definitely not my intent.

This was Eugen's intent from his post.

> I continue to believe reposts will lead to, on the whole, larger amounts
> of traffic reaching that site.  I don't believe this makes it legal.
> It's an orthogonal consideration.

This presumes that you are in fact increasing traffic to the site, when 
by the very act of reproducing the content, you *reduce* the likely 
traffic to the site *because* you reproduced the article. That is, the 
theory behind "hey, reposting generates more business" is exactly 
identical to "hey, piracy increases software revenues." Both are at best 
fallacious arguments, and more empirically, not supported by the evidence.

In fact the evidence from other industries runs entirely counter to 
this.  I'd post a specific link, but it might be banned by various 
filters.  Email me offline if you want it.

What it comes down to is, whether you pay for it out of your pocket or 
via advertising, or not, the "or not" wins, and causes hardship for the 
content producers/distributors.

Empirical evidence vs a WAG based upon an argument that is built on 
incorrect assumptions.

>>> any form or by any means..."  So I think the case that snippets are ok
>>> but full reproductions are not is a difficult case to make.  Perhaps
>>> there are laws that preempt this that discuss what percentage of a work
>>> can be reproduced without it infringing on copyright that I'm not aware of.
>> In the US there is a concept of "fair use" which allows some snipping.
>> The courts have been arguing over what fraction constitutes fair use.
>> Pretty much all of them are in agreement that 100% does not fair use
>> make.  That has gone over rather poorly for those replicating 100% of
>> others work.
> Agreed, and thanks for the note about "fair use."  That is indeed the
> part of law I was referring to but had managed to forget exactly what it
> was called.  Interestingly, a quick google search gets me to:
> http://www.copyright.gov/fls/fl102.html
> This page has four factors that are considered, and since two of them
> would be in favor of how Eugen shares full posts, whereas two of them
> would be against him, I continue to be on the fence about what I suspect
> would actually happen.  I'm sure there's some case law that would
> clarify that.  Whatever the legal decision in the US, I doubt this has

Heh ... so you've not been paying attention to what the music/movie 
companies have been doing here in the US to people who want to 
redistribute content without being troubled to pay for it to others 
happily consuming it without paying for it?  How is this not identical 
to redistribution of an entire article, without the benefit of providing 
the consumer of the article an option to pay for their consumption via 

Really, there is no gray area on this.  I know people might like to 
believe this, but its pretty cut and dry.

In fact its so cut and dry that some folks have started suing over "deep 
linking", even though its precisely that which generates traffic to 
their sites, and allows them to collect that revenue. I'll classify such 
people doing that as ... not entirely understanding of the way the 
internet works.  But its those links that give companies the chance to 
generate revenue.  They made the decision to package these with their 
content, and it is not up to us as external actors to strip this out for 
everyone else.  We can decide to do this for ourselves (the analogue is 
the "mute" button on your TV remote, which allows you to avoid listening 
to commercials which pay for the TV programs you might watch, if you 
watch).  But there is no neighborhood-wide mute button, nor should there be.

> any practical ramifications for him (besides possible removal from the
> list if this is against policy).  It's not going to be worth it for any
> of the people Eugen reposts to try and sue him, particularly across
> national boundaries.

Both Eugen and the folks he is copying from are in the EU.  They won't 
go after him, his pockets are not deep enough (no offence to Eugen on this).

> Moving on from considering this from an academic perspective, I'm
> personally for snippets in terms of raw words in the email, but
> definitely more selected and/or larger snippets than the one small part
> of the first paragraph I get in my RSS feed from scalability.org :D.

I don't advertise on that site.   And Wordpress' RSS features are borked.

> Some of those are utterly useless in conveying what the article is
> really about.  I suspect this is my RSS reader or your RSS server
> though; no doing of your own.  I just mention it as an example of where
> a larger piece of text makes all the difference in the world in terms of
> my going to the site to read it or not.

Actually most of the traffic we are getting there comes from the Twitter 
reposts which provide a shortened URL back to the site.  Its hard to 
write meaningful posts in 140 characters or less.

>>> On etiquette -- I absolutely abhor advertisements, which is why I use
>>> AdBlock and FlashBlock on Firefox.  This is a HIGHLY ideological
>> Etiquette is not about advertising.  Etiquette is about not abusing a
>> resource, not blasting large messages which may be better read online
>> via pointers.
> I disagree.  If we can agree that this type of reproduction falls into a
> legal gray zone because it is for non-commercial and educational use, or

So ... reproducing a commercial work, which has a specific set of 
linkage options, but no reproduction options is legally gray because it 
is "for non-commercial and educational use " ?

By that logic, I can rent DVDs from Blockbuster, make copes for 
"non-commercial and educational use".  I can borrow books from my 
library, and make copies for "non-commercial and educational use". Ad 


That isn't a viable argument.  There is nothing even remotely gray about 
this.  If its a commercial work, in current circulation, how is 
distributing it, gratis, without compensation, permission, or the 
possibility of compensation, to hundreds or thousands of people "for 
non-commercial and educational use " ?

I've made copies of books from the library, when the source was out of 
print, I could not find another copy anywhere, and it was, indeed, for 
educational use (for my thesis research, where I needed to have a copy 
close to me to so that I didn't get delayed in my work).  I used that 
very sparingly, as the "non-commercial and educational use" is not 
absolute, and it would be unfair, be bad form, be unethical for me to 
make copies of "Feynman's lectures on Physics" and distribute them 
gratis to many students, "for non-commercial and educational use" versus 
pointing to the site where people could procure their own.  The only 
time I've used this is when the market no longer provided a copy, and I 
could not find one anywhere else, inclusive of the publisher refusing to 
look for or print a single copy.

Which in a different sense is why I like (sort of) e-books. Production 
costs don't include printing, so reproduction costs for works can be far 
lower, reducing that issue.

> at least can agree that the likelihood of Eugen being sued over it is
> very low, then all that is left is whether or not full reposts like his
> are respectful to the authors or not.  Since the only reason why people

It has nothing to do with respect.

> would be upset about the repost happening in full that I can think of is
> loss of advertisement and thereby revenue, this boils down to how full

Yes ...  which happens when you reproduce things gratis without the 
advertisements or permission to reproduce them.

> posts impact the traffic.  Ergo, etiquette is *only* about
> advertising/money.  As an example, I have NO problem with people

Um ... no.  Etiquette is about respect for others in a social context.  
C.f. http://en.wikipedia.org/wiki/Etiquette Its about behaviour and 
actions.   My argument is that it is a breach of etiquette for someone 
to force me to read something by sending it out in its entirety, versus 
sending me a link and a cogent reason to read it.  This is the case with 
huge articles in emails, or autoplaying flash sections on websites, 
etc.  Its a breach of etiquette to the community as a whole to impose 
your specific will (reading, playing, etc.) upon them without their consent.

> reposting my academic papers or posters.  The ACM or IEEE might,
> however.  I make no money from my works (at least directly).  They do.

Au contraire! Are you not paid for *producing* your works?  Would you 
not be at least mildly upset if you were asked to produce those works 
for less money?  Or no money at all?  I am talking about your salary 
here, not any additional money you might make from authorship.  You are 
paid for research and related work, correct? If you were paid less for 
this work, would you not be understandably upset?  That is the core of 
the argument as to why it is wrong to reproduce the work, as it lowers 
the effective income potential for the producers of the work.

There really is no gray area here.  It is quite cut and dry.

> So all that matters is the $s in terms of respecting or disrespecting
> somebody.

Again, no.  What matters is that someone who should not have the power 
to make decisions about whether another group has the right and ability 
to make money off of their work, has arbitrarily decided that they 
should be paid less.  The breach of etiquette is a separate but also 
important issue.  My argument is simply that there is strong empirical 
evidence from other ... er ... industries (think San Bernardino area, 
but there are many others as well, think Malibu and Studio City area) 
... that what is postulated (free content drives more traffic) is most 
assuredly not the case.

>> If you don't like advertisement, thats fine.  You don't have to.  You
>> don't have to click on it.  The media that covers us, and pays Nicole's,
>> Doug E's,  Rich B's and many other good folks salaries, are paid by
>> advertising.  So, does depriving them the chance at getting revenue for
>> their hard work (writing articles) by replicating their postings in
>> their entirety bereft of adverts here for all to see ... help them
>> continue to cover and write about this market?  That is, by doing so,
>> you have effectively deprived them of their ability to obtain
>> compensation for their work, in the form of advertising revenue.  Is
>> this right for *you* to decide to do this to them?  Would you feel
>> unhappy of *they* decided that you have to have your salary reduced by
>> some random amount, though you did the same work that usually results in
>> more salary for you?
> I'll continue your analogy here to demonstrate my point that this is in
> more of a gray area than you suggest:
> How would I feel, if somebody decided to increase my salary by some
> random amount, but in the process used some of my work in a way I didn't
> agree to?  I'd probably be inclined to agree to that use, even if that
> use wasn't fully optimal for myself.  It's a net-gain for me, so unless
> I felt I could get that optimal gain myself, without their help, I fail
> to see how this would truly upset them.

See above.  This isn't a gray area at all, and the assumption of 
increased traffic to the site is, in fact, well known not to be 
correct.  This is seen in pirated movies, books, etc.  Microsoft doesn't 
get more money from pirated copies of its OS/apps.  The film studios 
don't earn more money from bittorrent users.  NYTimes and others were 
not making any money from people reproducing their articles in whole.  
So what makes you think that the site in question would buck the trend 
over the last few decades and start making more money from 
redistribution of its content?

Lets stick with the reduction in salary argument, as this is what the 
evidence supports.  So lets randomly reduce your take home pay, by doing 
something you do not explicitly permit.  This decision is not made by 
you, but by a third party, outside of your employment, who simply 
decides you need to earn less money.  This is exactly what is going on 
here.  There is nothing gray at all.  Arguing that full reproduction 
increases traffic goes against the empirical evidence that it does 
precisely the opposite.  With that in mind, are you ok with some third 
party arbitrarily reducing your take home pay for reasons that are, 
well, laughable at best.

> Again, sure, if Eugen decides to post in the future just smaller
> snippets, he'll increase their "salaries" by larger random amounts, but
> ultimately I'll argue he is not beholden to do so (from a purely
> practical, non-legal perspective) if he doesn't wish to, nor is doing so
> truly disrespectful if we can agree he is increasing their traffic (and

... but he is decreasing their traffic by posting the *WHOLE* article.  
He *could* increase their traffic by posting a snip and a link.

It has nothing whatsoever to do with respect.   It has everything to do 
with rights of the content producer to make a living from their 
content.  It has everything to do with bad form by forcing content down 
list readers email clients without their consent to read it.

> thereby ad-serves).  This is particularly true because Eugen isn't
> taking some part of that ad-serve revenue for his own.  If that were the

... no, but he is reducing their advertising revenue by posting it in 
whole, as he is reducing their traffic.  There is no empirical evidence 
that his posts generate more traffic, there's a great deal of evidence 
across many content publication industries that it reduces traffic.

And again, as these are the folks that depend upon their content to 
generate revenue and continue to cover this area, would it not seem that 
a sensible approach would be to, I dunno, help them?

> case, it would be a far harder decision for the author.
> I agree that it's disrespectful if the link isn't along with the full
> text, or if the authorship is removed or buried.  That's clearly just
> stealing work and pretending it's yours, which is disrespectful for
> obviously plagiarism reasons.  Eugen didn't do that at all.

Uhhh ... so you are ok with ripping a DVD and distributing it, as long 
as you are not claiming it as your own work of authorship?

Seriously, this is a flawed argument, at its most profound and 
fundamental level.  The injury occurs the moment a decision is made by 
someone whom doesn't have ownership or decision making authority to make 
the decision to distribute without support for advertising. Everything 
beyond that is moot.  Once that decision is made, game over.  The injury 
has occurred.  Arguing that he isn't plagiarizing it is superfluous.  
The damage isn't there, its in the redistribution without any hope of 
compensation due to a decision made by the redistributor to consciously 
remove the means for that compensation.

>> There really isn't any gray area here.  Its pretty much black and white.
>>     The Inquirer, The Register, HPCWire, InsideHPC, ClusterMonkey all
>> depend upon people clicking those links for them to make money.  Is it
>> your right to decide that they cannot make as much money by full
>> reproduction to a wide group of people?  Or would you in fact do them a
>> favor by "driving traffic to them" which, once you start processing
>> this, you realize is silly.  You've consumed an advert supported product
>> delivered to you without adverts, and now you are going to go to the
>> site to see adverts?  Um ... no.
> You're failing to see that many of us aren't actively RSS'd to many of
> those sites.  So the potential to get those of us who aren't served ads

Doesn't matter if you RSS to those sites.

> to, was 0%.  Now it is somewhat larger than that.  Even if it is 1%,

Did you read the article in whole on the email, or did you immediately 
visit the site when the article showed up?  Most folks would do the 
former.  Very few, if any, the latter.  Some might have done this after 
I raised the point.

I clicked the link and read it there.

> that's a net-gain.  I went to the site because my email is on my left

No ... it is a net loss.  Again, content industry after content industry 
has struggled hard to deal with this problem.  And it is a problem for 
them (for their business models).

> screen in my triple-monitor setup, so reading the whole article in
> Thunderbird was not an attractive proposition.  I much preferred to read
> it, and see any graphics along with it, in Firefox, which was on my main
> larger screen.  Though I didn't see them, I'm sure they served me some ads.
>> The issue is if the Inquirer folks call up Penguin and say "guys, stop
>> this person from posting our articles without permission, and remove all
>> the old ones he's posted", what do you think is going to happen?
> I continue to agree with everybody that some policy should be shared
> with the list regarding how much is acceptable to snippet or copy from a
> copyrighted site.  I don't think any of us want to get Penguin in hot
> water, whom we all appreciate providing this service.

Penguin are good people.

No more long articles.  A snip and a link.  Please, it is better for 
everyone, and you might actually generate more traffic to the site in 

If you make it so that people cannot make money in an industry, they'll 
go do something else.  This is not what we want for HPC journalism.   
Especially since we have quite good journalists in this space.

Its a choice to make, and you get to choose by your actions.  My point 
being that what James Cuff did a few hours ago with a link was exactly 
what should have happened.  What Eugen did is what content producers are 
fighting against, as it directly threatens their business models.

Joseph Landman, Ph.D
Founder and CEO
Scalable Informatics, Inc.
email: landman at scalableinformatics.com
web  : http://scalableinformatics.com
phone: +1 734 786 8423 x121
fax  : +1 866 888 3112
cell : +1 734 612 4615

More information about the Beowulf mailing list