There are systems that should have extremely tight specifications for
everything, and it does not matter how much CPU time is spent on any
portion. You have worked on some of them, congratulations.
Now, you need to shift gears. The credit system in BOINC is NOT one of
those systems where the amount of CPU power thrown at it is irrelevant.
The validation system in place will catch all but systematic errors (where
either there is a widespread hardware malfunction like the Intel floating
point error of a decade ago, or a mistake in the software). The mistakes
in the software can only be detected by normal software engineering means.
The widespread hardware malfunctions can be detected by running a reference
task once and comparing the results of that to a reference result from a
reference machine.
BTW, many people here also have excellent qualifications, myself among
them.
jm7
"Paul D. Buck"
<p.d.b...@comcast
.net> To
"Lynn W. Taylor" <l...@buscom.net>
09/29/2009 08:33 cc
AM john.mcl...@sybase.com, BOINC
Developers Mailing List
<boinc_dev@ssl.berkeley.edu>
Subject
Re: [boinc_dev] [boinc_alpha] Card
Gflops in BOINC 6.10
On Sep 28, 2009, at 5:30 PM, Lynn W. Taylor wrote:
> Because I've watched you, Paul, and your accusations and your
> personal attacks and your constant complaints that the BOINC world
> would be much better off if the world would just adopt your grand
> schemes.
I have been thinking about what this statement implies. Especially
the part about "grand schemes" and all that that kind of reference
implies ... And I am going to try to provide an answer, not that it is
likely to do any good ...
In 1970 I entered the US Navy while I served my 20 year career I
achieved the rank of Senior Chief. My largest normal span of control
was 43 sailors.
I also attended 399 days of school (weekend not included) in 30
different subjects (school were from 1 to 130 days) the longest was
the equivalent to an AS/BS in electrical engineering (less English and
such).
I have also completed 16 commercial correspondence courses
concentrating in Computer hardware and software, 123 navy
correspondence courses ranging from hand tools to military law (GPA
3.82), 11 Army correspondence courses (I learned how to use nuclear
weapons on bridges and other fun topics including a lot more than I
really wanted to know about Nike/Hercules missiles), awards and
decorations are not terribly relevant so we shall skip them, but in
the 70's and 80's I wrote several thousand deficiency reports on
documentation and the software in use on the computer driven test
systems i worked with. I attended conferences where the technical
content of of documentation was reviewed and several navy wide
technical investigation boards. During the majority of my naval
career I subscribed to in excess of 100 technical publications a
month, all of which I read (sea duty ... gotta love that spare time
with no distractions) almost all of which related to systems, systems
engineering, electronics, computers, programming languages, and
computer hobbyist's interests (I only missed the first three issues of
Byte and my first computer I built out of a PC board in a baggy,
mimeographed instructions and my soldering iron)...
I have a Bachelor of Science in Computer Science (Summa Cum Laude) and
a Masters in Software Engineering.
I worked on the OTH-B radar software (~3.5M lines of FORTRAN) and
another classified program reverse engineering a phased array for the
US Air Force as Civil Service (GS-12 with no bonus from military
service)
At SAIC I was the technical lead (read: lead designer) on teams of
4-15 people developing Relational Database systems from the database
to the UI front end using Oracle and PowerBuilder. When I left the
company I had performance options in excess of mid 6 figures.
I have taught ~30 different classes/courses at the university level at
3 different universities in computer science and Oracle Databases (the
official Oracle courses at UC Davis Extension).
I started in SETI Classic in July of 2000 and though I cannot pin it
down I was in mid-BOINC Beta when it was still closed (~2004?). I took
on the task of converting the technical documentation of SETI Classic
which after 4 generations evolved into the UBW ...
Over my life I have programmed in 6502 assembly, FORTRAN, Basic,
Pascal, Forth, Vital, 8086 Macro Assembler, PowerBuilder, PL/SQL, SQL
and I have familiarity and can read C/C++ and at least a dozen other
languages.
I am currently disabled and homebound which means I have lots of time
to watch BOINC and to think ...
So, why would I tell you all that? ... because your statement "Grand
Schemes" has the implications that I have neither the technical
training nor the experience to make evaluations and from those
observations to make recommendations about what is going on in the
BOINC System and how to make it better. In BOINC Beta and initial
deployment not only was I the sole person documenting BOINC but I was
also one of the 4-5 people that patrolled the boards back then and
attempted to help those in need (go back in my SaH account and read
the messages, they are still there). I have run BOINC on as many as 9
but never fewer than 3 systems (sole exception is a 2 year sabbatical
for health reasons). I usually have one up and running and I watch it
all day long ... my systems have tended to be on the leading edge of
computing capability which is why my world position is roughly 180
(peak 141) ... for a person on disability and with only access to
computers I own that is not too shabby ... (currently 36 CPU cores and
9 GPU cores)
So, your statement has the implication that I don't have the right by
your lights to make suggestions... I suppose that might be fair
enough ... but before you use careless language that implies, as you
did, that a person is not competent to make the evaluations that I
have, or to try to make the system better by offering design changes,
be sure of how much you know about them ... and not just what the
rumor mill might be providing you in the background ... the reasons
that many of the proposals I have made over the years may seem, ahem,
grand is that some of the problems that no one seems to want to
address are also equally "grande" ... And if you evaluate my
technical proposals based on my interpersonal skills then your
criteria from making an evaluation of them is far more flawed than my
inability to communicate.
There is a middle point that is interesting because, though you have
not appeared to notice, I am also all in favor of what Martin is
proposing. I suppose that does not count because no one seems to be
interested in his ideas either ... grand or not ... then again, I have
been proposing the kind of overhaul to the credit system he is working
on since before BOINC Beta opened to the public ... I have also
endorsed ideas John has proposed in the past though in recent years he
is mostly saying do nothing as a solution to all known problems ...
and when anyone makes a suggestion that makes sense I support that ...
when they make a suggestion that I think could be improved, yes, I
offer criticism ... and when I think a suggestion is bad I also say
why I think so ...
The second to the last point is more implied than direct ... it is
that a vigorous defense of a person's proposal by the person that
proposed it is somehow wrong. Yes I defend my ideas as I would expect
you to defend yours and John to defend his ... and if we stick to
facts and have the spirit that we can find common ground then, yes,
progress can be made. But progress cannot be made if the alternative
solution offered in rebuttal to any of the issues raised is to do
nothing ... which lately, more than ever before, seems to be the
default reaction to change proposals by just about everyone ... we
will stay in the same bad places we are in now ...
As to personal attacks? Again I find that interesting because aside
from asking you to abide by the rules you wanted to apply to me the
only other things that I can recall that might be construed as an
attack were the observation that I made to John that he did not read
what I wrote, if needs be I will go back and find the message where he
made that admission that he had not. Thus, not an attack but an
accurate observation. The only other thing I can think of is when I
suggested that his rebuttals sounded hysterical. Understand that I
have read everything John has written that I am aware of... we have
exchanged e-mails and engaged in long discussions over many years now,
I am not sure but I think we even talked on the phone though I cannot
swear to that ... so I think I know the tone and tenor of his
writing ... when he continues to repeat the same rebuttal over and
over even though I have altered the suggestion or proposal to
accommodate his objection; well I also think that that is a fair
observation ... perhaps unkind, and maybe unnecessary, but not untrue,
and not an attack ... at worst it is one man's opinion on another
person that they have known a long time ...
The bottom line is that if you think that I am arrogant to think that
I know what I am doing, well, good enough ... but, I do have the
training, the credentials, and the experience ... and yes, I think
that those do give me a little bit of a right to make observations and
critiques of The BOINC System. When I suggested that the BONC
database is a poor design it was because that is what I did and what I
taught for a living. Even better I also consulted (because I was
still working at the time) with DBAs that had been doing databases for
decades more than I ... when I make other observations and
recommendations it is based on years of experience with computer
systems, research, thinking, and observation of the way the BOINC
System works ... which based just on productivity, well, I beat almost
everyone in the world in accomplishing work with BOINC ...
Compare and contrast:
http://boincstats.com/search/all_projects.php?cpid=a6477942e70ed39f669d1ff2ede05be8
http://boincstats.com/search/all_projects.php?cpid=a0d4f766815daaa87fdcfafcbed5c8de
Is it a "fair" comparison? Maybe not, but it does mean, if nothing
else that I have seen more work flow across my systems that John
has ... is that meaningless ... you can say that I suppose ... but it
still means that it is far more likely that I have seen more tasks
processed then John has ... of course, I am sure you can also twist
this to be an attack rather than a fact based observation ...
Anyway, ... Yes, Virginia, Paul does have a crazy notion that school
learning in a discipline and 34 years of real-live experience does
give him some standing ...
_______________________________________________
boinc_dev mailing list
boinc_dev@ssl.berkeley.edu
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.
