[nylug-talk] Continually discrediting oneself -- WAS: Slim home server for samba and subversion ( and possibly IMAP)

[Bryan J. Smith]

I almost resisted respond to this, as I have several others.  I thought
long and hard how I should respond, if I did.  I do this fully knowing I
should not.

Alex Pilosov wrote:
> a) IXP is expensive, nobody puts IXP on a 200$ NAS device (with rare
> exceptions, iomega did I think)

Chris Knadle wrote:
> The $90 Linksys NSLU2, runs an IXP processor.  [I'm using two of these
> with Linux on them for DHCP services.]
>      $ cat /proc/cpuinfo
>      Processor       : XScale-IXP42x Family rev 1 (v5b)
>      BogoMIPS        : 131.48

Chris --

Sometimes the best response is not to make one.  Although some members
won't know what we are talking about, or may believe Alex's response to
mine, those in the industry will know.

I.e., Alex has once again discredited himself as being someone who has
any experience in an area -- namely exposure to one of the most
proliferated embedded solutions for SOHO networking/storage devices,
Intel IXP400 series.

E.g., from this, he has shown he cannot separate the cost of a prototype
board from the per-unit quantity cost of mass produced,
application-specific device with the same microcontroller -- again,
namely the Intel IXP400 series.

Furthermore, he goes on to make such broad, non-specific statements that
vary greatly from application to application, like ...

Alex Pilosov wrote:
> b) ARM has enough firepower to serve out a gigabit of data

and ... (several times) ...

Alex Pilosov wrote:
> it's a software issue.

I have often found myself keeping my mouth shut throughout my career
with select CS and IT folk have lectured me about "digital" and
"software."  I distinctly being in the library (waiting on a resource)
at my Alma Mater when a CS major (in front of me, using the resource)
lectured me about architecture, and cut me off a few times when I tried
to lead him towards where he should be going.

There are major interconnect, buffering, ASIC and other considerations
when it comes to any I/O, especially multiple dataflows through I/O,
possibly relying on LOAD/EXEC/STOR and other operations.  The second I
start talking about network and storage processing engines, there's
always someone trying to "look smart" with "big O" and other things.
Intel still designs specialized ASICs around their cores, even the new,
2-issue, in-order x86 Atom design.

Lastly, There is a huge lineage of ARM products.  Saying ARM is like
saying x86.  It says nothing but a huge lineage.

The continual lack of specifics in his responses, and ignoring my
further engagements to draw out his thinking or more info, is beyond
being a troll.  I not only look back at my youth, but I regularly have
to keep myself in check, to differentiate between when I'm being
intuitive, inquisitive, complementary, confrontational, etc... and when
I'm just downright insolent.  But in all honesty, I don't think I've
ever seen this continually unique combination of insolence with
providing statements that just flat-out discredit oneself.

I'm not big on credentials, and it takes _years_ to earn the
professional trust of others.  I fully expect people to not believe a
word out of my mouth -- until I prove, over _years_, that what I say
tends to provide them with answers that work for them.  Only then will
people stop to listen to what I say, and not before.  And even then I
hope they will still question and test things for themselves to ensure
not merely that I wasn't mistaken, but my experience and application was
also applicable to their need -- which is not always the case.

Alex, you are assuming beyond belief, and just utterly discredit
yourself.  It's clear you railroad over people, not merely making
assumptions about where they are coming from, but clearly exposing you
have not been exposed yourself to many industries, where experience
would have informed you of things to the contrary.  It's really hurting
you.

I mean ...

I could bring in an engineer who has designed the reference firmware and
software of some of these SOHO devices used in every "household name"
products on the shelf at superstores, someone who has worked as a core
FAE at Intel's primary partner on IXP designs, and you'd not only argue
with him -- but attempt to discredit anything he posted just trying to
help others, and make every attempt to do so whenever he posted at all.

I could bring in an engineer who has worked with Steve Furber (I fully
expect you to Google him now, and then act like you knew -- I'll save
you the trouble, you'll find him on the "ARM" page at Wikipedia to
start ;), or been on his teams at educational and industry leaders where
he is (or was) on the board overseeing development -- and yet you'd
lecture him on ARM anytime X-Scale was brought up, or make qip's about
ALUs and ignore his focus on ASICs when he tried to steer you.

I could bring in an electrical engineer who has helped define several
IEEE 802 sub-committee standards and revisions, and you'd lecture him
it's all about software.  You assume everything, and will not merely not
listen to no one, but you will not even work with someone who honestly
tries to lead you in the right direction.

I know it was wrong to post this, but I hope you see it Alex.  But my
initials are "BS" and I even put "Professional, Technical Annoyance" in
my signature.  That could mean I'm all BS and should be ignored.  Or it
could mean I have learned some humility in my time -- hint, hint.  ;)

-- 
Bryan J  Smith              Professional, Technical Annoyance
mailto:b.j.smith at ieee.org  http://www.linkedin.com/in/bjsmith
-------------------------------------------------------------
           Fission Power:  An Inconvenient Solution