Gecko Hamaker

Roger,

I don't know where we are going to end up with this
thread. But I
know that RIGHT NOW - we do not have multifiles inside
Gecko. The
students never got this far. Most of your great
results obtained
thru grams involved using multifiles - i.e. varying
EMA, etc.

So I am proposing that implementing multifiles might be
an immediate need.

Lin

At 08:03 PM 8/1/2006, Roger H French wrote:
>This is a great thread,
>
> I think that the way we need to go in general,
is to run projects
>that have internal variables, and give a series of
A's, where we can then
>see how A varies with a certain change.
>
> Gecko Hamaker does one at a time because it
needed to start
>somewhere, and it introduced projects, to help group work.
>
> But I've been liking the case of making
projects with interpolation
>layers where we can vary composition (ema layer),
composition among layers
>(coupled mixture layer), the gradient layers, etc.
>
> So the ema layer is aversion of making a bunch
of JCV's and then
>getting out a multifile.
>
>NOW for Rick's CNTs
> I'm not saying we need to do it in Gecko, you
can do you perl
>scripts, if that gets your work done.
>
> But as Lin said, "The "work" in Gecko falls
into two categories: (1)
>adding materials into the database, and (2) creating a
project."
>
> I think that we need to allow multiple file
processing in (1), so
>that we can add a bunch of spectra, and also calculate
their London
>Dispersion Spectra.
>
> But for the projects, they are also stored in
our MySQL database.
>And maybe we should think of them as a way to do what
I'll call "Extrinsic
>Variable Problems". The A(L) is a variational
problem, but insights come
>from looking at how A(L) changes with changes in
various "Extrinsic
>Variables" which are fixed in any particular calculation.
>
>A vs different CNT LD spectra
> Rick wants to do a simple (or multilayer
configuration) and just swap
>in a lot of different LDspectra. One for each of his
CNT's and to probe
>the variations of A with CNT spectra.
> For me I think that the mysql script to
just modify the project
>to swap in spectra already in the database sounds like
a great idea. and
>so for a particular layer we just have it be
associated to a series of
>check boxed spectra in the database, and this defines
a family of projects.
>I think of this in the GUI, but it might (and most
probably is) easier in a
>simple MySQL Script). Now for a project "pack" like
this, I coudl see
>having the A spectra going into a simple multifile
spc. (I am still
>working to get spec view to let us see the multifile
spc's) I think that
>having the spectral A's come out as a multfile helps
see the what the data
>is telling us.
> Rick, when you calc the A's with your own code,
are you doing on 123
>type problems and getting only non retarded results?
cause I find the A
>versus L curves hold alot of suprises.
>
>A vs layer thicknesses in a multilayer problem
> We always calculate A vs L where L is the
change in thickness of the
>"central" or variational layer. But in mulitlayers, I
want a way to
>calculate for a bunch of problems where the other
(fixed) layers have their
>thicknesses chagne. and make a multifiel and let me
scan how the problem
>changes.
>
>
>So I don't know where we go for introducing these
extra levels of
>variation, but a series of projects or a mysql
scripting could really open
>up the problems we can study.
>
>
>
> Regards,
> Roger
>
>--------------
>Roger H. French
>DuPont Co. Central Research
>E356-384, Exp. Sta.
>Wilmington DE, 19880-0356
>Phone: 302-695-1319
>Mobile: 302-983-3003, AIM: vuvrhf
>email: roger.h.french@usa.dupont.com
>
>
>
> ldenoyer@spectrum
> square.com
>
To
> 08/01/2006 04:15 Richard F
Rajter <rick_14@mit.edu>
> PM
cc
> Roger H
French/AE/DuPont@DuPont
>
Subject
> Re: Gecko,
visit, energy wings
>
>
>
>
>
>
>
>
>
>
>Rick,
>
>Thinking about your need to automate. I am just
wondering: why don't you
>automate the creation of projects through mysql?
Wouldn't that be easier?
>Then you would have 100 new projects in your database,
and you would just
>load them into Gecko one at a time and hit the
calculate button.
>
>The "work" in Gecko falls into two categories: (1)
adding materials into
>the database, and (2) creating a project. In both
cases, I am thinking
>that writing a mysql script to get both the materials
and the projects into
>a database is the way to go.
>
>Lin
>
>At 09:57 AM 8/1/2006, Richard F Rajter wrote:
> That is correct. You would have a "project
file" that you'd pass to
> Gecko and
> say "calculate".
>
> Here is the benefit.
>
> I can take 1 project file as a template... and
create 100 project
> files all
> with a different "layer 1 spec". I can then
use this same script to
> pass each
> one to Gecko in a row. All without the need of
going through each
> and every
> one.
>
> I'm a data monster... unlike Roger who is
calculating one thing at a
> time. So
> I'll have to explore a bit and see if this is
possible without any
> major
> changes. I'm assuming it will require
reasonable changes, so I'm not
> too
> optimistic at the moment.
>
> -RR
>
>
>
> Quoting Lin DeNoyer <ldenoyer@spectrumsquare.com>:
>
> Rick,
>
> How are you going to set up these project
files that you pass
> on the command line? Thru Gecko? If so,
isn't that exactly
> the idea of having different projects?
You set up each one
> individually, then save it? No swapping
of files each time
> you need to recalculate?
>
> What am I missing?
>
> Lin
>
> At 08:30 AM 8/1/2006, you wrote:
> Roger.
>
> Yes, being able to add LOTS of
spectra all at once would
> be awesome.
> But at the
> moment, it is file by file and with
60 tubes x3
> directions... you can see how
> that would get a wee bit "tedious" :)
>
> But yes, looking forward, it only
has to be done once and
> that's that.
>
> What I'm actually most interested
in, and I'd have to run
> this by Lin to see if
> it's possible in the given
framework... is to be able to
> be able to "pass" the
> project file on the command line
and have it calculate it
> that way.
>
> The advantage would be the
following. Suppose I wanted
> to calculate the coated
> substrate coated nanotube system at
10nm away for several
> CNTs. I could do so
> manually one by one, merely
switching out the tubes. Or
> I could conceivably
> pass it a series of project files
with the CNT properties
> switched in each one.
> They would calculate and dump to
the save directory...
> to be harvested later.
>
> I realize for the work done so far,
this ability to
> calculate huge sums of data
> was not really necessary because
the number of components
> were not altered in
> any significant way. So doing such
a thing would have
> been way overkill. However, for MY
purposes, it is not
> overkill at all...
> and almost becomes a
> necessity to really be able to dig
through the data.
>
>
> That is why I've done it by simple
perl scripts for
> now... it's certainly not
> ideal, but I can chug through an
A123 calc for CNT1 water
> CNT2 in a 180X180
> grid in less than 5 minutes.
>
> Would I like to do this in Gecko?
Absolutely, when it
> can be passed to the
> engine via the command line... so
it would be equally as
> trivial and I would
> gain access to the add-a-layer
engine. But this will
> certainly take time to
> implement. It's not going to be a
1 day job as far as I
> can tell.
>
> -Rick
>
>
> -Rick
>
>
>
>
>
> Quoting Roger H French
<Roger.H.French@USA.dupont.com>:
>
> Hi Rick,
>
> your question
> "I had a question on how you
implement the "high
> and low energy wings" during
the e2 to LD
> calculation in GRAMS/Gecko
> Hamaker.
> The reason is, it is far more
efficient for me to
> use my own short program
> and
> script this for the 100+ e2
curves I have from
> Wai-Yim, than it is to go
> through one by one clicking
all the right buttons."
>
> The exponents of the wings
vary depednign on
> whether you are doing the
> london dispersion transform
of the epsilon 2 file
> or the Jcv file. Lin
> can give us the correct high
and low energy wing
> exponents for epsilong2
> and jcv1. Beware that I
think in our silica prb
> paper, we listed the wing
> powers incorrectly.
>
> But i would think that
instead of even
> scripting 100+ london
> dispersion transforms, and
continuing to do this in
> your standalone
> hamaker calculater, which
won't get used intot he
> future. Isn't it just
> easier, and leads to an
advancement for all future
> users, to implement two
> things. One in the add
spectrum to database, that
> one can not just choose
> a single file name, but
instead choose a bunch of
> spectra to import.
>
> And similarly for the
LD transform, one can
> check the box on a
> number of spectra, and then
have a bunch of them
> get the london dispersion
> spectrum calculated.
>
> I've always found its
better to just do
> things in a way that gives
> a permanent capability, your
frustrations on
> multifile import and
> transform will persist, even
after you do the cnt's
> spectra.
>
> Lin, whats needed to
allow mulitple spectra
> import and LD
> transforms. Your in the java
at present, shall we
> do that?
>
>
>
> Regards,
> Roger
>
> --------------
> Roger H. French
> DuPont Co. Central Research
> E356-384, Exp. Sta.
> Wilmington DE, 19880-0356
> Phone: 302-695-1319
> Mobile: 302-983-3003, AIM: vuvrhf
> email:
roger.h.french@usa.dupont.com
>
>
>
> Richard F Rajter
<rick_14@mit.edu>
> 07/31/2006 10:46 AM
>
> To
> Roger H French/AE/DuPont@DuPont
> cc
> ldenoyer
<ldenoyer@spectrumsquare.com>
> Subject
> Gecko, visit, energy wings
>
>
>
>
>
>
> Roger,
>
> I'm back at Dupont for 2
weeks. I had a question
> on how you implement the
> "high
> and low energy wings" during
the e2 to LD
> calculation in GRAMS/Gecko
> Hamaker.
> The reason is, it is far more
efficient for me to
> use my own short program
> and
> script this for the 100+ e2
curves I have from
> Wai-Yim, than it is to go
> through one by one clicking
all the right buttons.
>
> Lin + Roger,
>
> I meant to do this last week,
but I believe I can
> create all 3 install4J
> Gecko
> installers on my Mac. I will
use the 1.81 code Lin
> just updated onto
> SourceForge and generate all
3 tonight. I'll then
> try them tomorrow
> morning on
> my mac, one of steve's
windows machines, and
> Roger's linux machine to see
> if I,
> in fact, get all 3 working.
>
> Which would rock.
>
> -RR
>
>
>
>
>
>
>Lin DeNoyer, PhD
>Spectrum Square Associates, Inc.
>755 Snyder Hill Road
>Ithaca, NY 14850-8708, USA
>Website: http://www.spectrumsquare.com/
>Email: ldenoyer@spectrumsquare.com
>Phone: 1-607-272-2352
>Fax: 1-607-330-4750
>
>
=====================================================================
>
> This communication is for use by the intended
recipient and contains
> information that may be privileged,
confidential, or copyrighted under
>
> applicable law. If the reader of this message is
not the intended
> recipient, or an employee or agent responsible
for delivering this
> message to the intended recipient, you are
hereby notified that any
> reading, dissemination, distribution, copying,
or other use of this
> communication or any of its attachments is
strictly prohibited. If
> you have received this communication in error,
please notify the
> sender immediately by replying to this message
and deleting this
> message, any attachments, and all copies and
backups from your
> computer.