Gilles,

Thanks for presenting this useful information to ISOGEOCHEM readers, but
should it also be mentioned or is it fair to say that a substantial portion
of your salary is indirectly or directly provided by Isomass Scientific,
the Canadian distributor for Thermo Electron (formerly Thermo Finnigan)?

Although ISODAT 2 is 20 000 per mill better than "old" ISODAT, thanks to
the dedication, hard work, and expertise of Holger Jeglinski in Bremen, as
far as I can see, none of your observations apply to why Kinga Revesz
trashed a computer with the Version Handler of ISODAT 2.0, which she
described on ISOGEOCHEM December 1, 2004.  I estimate that this catastrophe
cost this USGS project of the order of 10 000 USD in lost research
productivity.

In fairness to GV Instruments, it should be mentioned that the MassLynx
data acquisition and control software is built upon Microsoft Excel and
Access, which are relatively rock solid.  Thus, it is no surprise to me
that Len Wassenaar has not had any difficulties in over a year.

Ty



                      Gilles St-Jean
                      <[log in to unmask]        To:       [log in to unmask]
                      CA>                      cc:
                      Sent by: Stable          Subject:  Isodat NT crashes, myths and reality
                      Isotope
                      Geochemistry
                      <[log in to unmask]
                      UVM.EDU>


                      02/03/2005 10:39
                      AM
                      Please respond to
                      Stable Isotope
                      Geochemistry






Dear all
Isodat NT: myths and reality
Version Handler (a later e-mail)

It would seem by the latest exchanges that Isodat is getting a little too
much of a bum rap.  Here I hope to clarify some of the myths about it and
the reality in its implementation.   The best way to do this is to explain
it through practice.  I will give a few examples of resolved problems
followed by each one of our own systems here at the Ottawa U lab (hardware
and software) where, perhaps, you can then judge your own systems from this
to determine the problem.

I have found that 9 times out of 10, the problem is a bad installation of
Windows or Isodat.  Other times it was either attempts at more advanced
features of Isodat without a good way out, if a major screw up occurred, or
a genuine bug.  I have accomplished all of these at one time or other and
have learned a great deal about computers and software in general.  Most
users are not particularly comfortable with computers or are not computer
literate, so a Windows crash would be indistinguishable from an Isodat
crash since it "was running" during the crash.  A defective modem, video
card or their respective drivers can sometimes be the cause (see examples
below).

Now we have had our share of Isodat NT gripes in the past, at times nasty
I'll admit, and still do on some aspects, but in the last 2 years the
software has been performing better and better with each passing month.  We
are now well beyond the "it does the job" to "it does the job really well".
Our lab has now reached a level comfort with Isodat NT that we stopped
printing the data on paper two years ago and archive all in digital form on
servers and CD backups.  We save $500-800 per year in ink and paper alone,
not to mention the storage problem associated with well over 15 thousand
analyses per year

I have asked, from Finnigan sales, how many instruments are, or can be,
running on the new platform (Isodat NT).  At present it is at least 1000.
Add to that 8-10 IRMS shipments per month and... you get the picture, a
"shitload" (MS Word's dictionary proposed "shipload", oh well...) of users
worldwide where each instrument can have more then one user (students,
scientists etc).  Thus the more users, the higher likelihood of having
novice to advanced users all trying different things with varying degrees
of success. The novices are the larger group today and this is from my
experience of training over 130 users in the last 4 years.  This leaves
close to 900 people whom either don't need training or can't afford the
time or money for it.  This problem is true for all manufacturers.  Finding
IRMS operators that can handle all is next to impossible for new labs where
they either can't afford or find them.

When a computer enters a facility it becomes part of the rules and
regulations of the Computer Services and Guidelines of the institution.
For example getting Windows XP SP1a (Service Pack 1a) was a problem for
some a few years ago because of these rules affecting access or deployment.
A subsequent patch following SP1a resolved many third party communication
problems such as COM ports and USB devices.  This had significant impact on
Isodat at the time for Autosampler and Peripheral communication (COM
ports).  Now the same thing is occurring with the deployment of Windows XP
SP2 in organisations thus causing extra problems with well written 3rd
party device drivers but with bad Windows handling of them.  I have seen
many installations of Finnigan competitors software as well and Windows is
not protected or updated by any mechanism other then the knowledge of the
user.  Perhaps as Dr Bao put so well, that the masses using the same
software may be larger on the Finnigan side !
 thus more potential for gripe.  Or are we just more vocal than the rest?
Mind you there is also another way of looking at it.  If out of 1000+
systems only a handful are complaining, maybe the rest are happy with the
work getting done.  We always hear the complaints and rarely the successes;
except maybe at conferences and workshops.  Human nature I guess... I do it
too.  When someone tries something and fails, we hear about it.  If they
succeed they get the work done and we don't hear a peep.  I'm sure there
are way more successes than failures with Isodat software.

-----------------------------
Examples

Example 1: At one site the user was complaining of the computer locking up
every morning and having to be hard booted.  Isodat was blamed.  Turned out
to be the "Sleep mode" in Windows Power Management was turned on without a
way to wake-up including the fact that the settings on the motherboard BIOS
did not like it either.  I deactivated it and the problem was gone.  This
kind of problem can also completely cut communication to a peripheral (GC,
Autosampler etc.)  In the same lab the computer locked up at random when
the TOC Analyser software was loaded in the same machine.  MacAfee
anti-virus was the culprit.  Got rid of that and the machine ran well for
the following three days that I was there.

Example 2: At another site the computer was infected with 26 virus and
worms, no anti-virus and free flow of diskettes from Isodat to their desks
to process further in Excel.  Needless to say crashes were regular and the
system sluggish, not to mention their desk computers... yech!   Who knows
how many of those are around?

Example3: I know of many labs that are still running initial versions of
Isodat NT from 4 to 5 years ago.  They have learned to work with the quirks
and prefer to leave it as is since it does the job.  The latest version
sits in a desk because they refuse to make the changes or are unable to do
the Windows-Isodat upgrades from NT4 to XP.  However when a problem occurs
some get pissed off that no help is available for that version and thus
gripes on Isogeochem making it sound like Isodat NT is nothing but trouble.
At least when a better version is available Finnigan tried to make amends
by sending out the CDs of the next version to all users of the latest
software... FREE.  If the users don't install it, then how far does the
hand holding have to go.  This goes for all instrumentation companies, they
are selling machines in the hundreds per year, not Microsoft's hundreds of
millions.  At some point the user has to take some responsibility in the
infrastructure and maintaining!
  of it.  The patches are available to all users on the Finnigan web site:
http://www.thermo.com/com/cda/category/category_lp/0,,15410,00.html
Look under Fast Links and click on the CIS (Customer Information Services)
link to get the latest manuals, patches etc.  Note you may need to get
registered where you will get a username and password

Example 4: Regular crashes due to bad drivers of computer devices.  A major
culprit in a few cases I have seen is the MODEM drivers.  They tend to
interfere with COM port communications.  Some are just flaky...period.
Make sure you have the latest drivers from the manufacturers of the
motherboard or devices (video, audio, modem, Ethernet etc.)  If you don't
know how to install them, find someone who does or pay someone to do it.
If you can spend $150 000 to $700 000 on the IRMS system, you can spend
$200 for a computer specialist to properly install the operating system and
drivers.  When a crash occurs a log is created which sometimes can point to
the culprit.  If you don't know how to access it or read it, get an IT guy
from your institution to help.  Most of the time the problem is in the
initial installation of the system or a corrupted system by worms or
viruses.  On rare occasions it is a piece of hardware that has failed.  For
example it took me 3 days to find an in!
 termittent problem with a motherboard's on-board harddrive IDE controller
that was writing garbage to the hardrive at different times.  Thus it was
corrupting the data and eventually the whole system was breaking at the
seams.  It gave the impression that the HD was failing.  Three HD later it
was finally pointing to the motherboard controller.  Replaced the
motherboard and never had a problem since (4 years ago).

In short, we feel very confident in Isodat, our experience has been trying
at first but the system has grown into reliable software and is improving
with every version.  However, a proper installation on a proper system is
critical.

------------------------------------------
OUR LAB WITH CRASH AND HANGUP STATS INFO

This info is from the users of the lab; including lab manager Paul
Middlestead, technical staff and students.
Hardware common to all 5 instruments:
All systems are running Windows XP French version w/Isodat 2.0.(SP from
1.24 to 2.34 over time), Windows XP SP2, Norton Anti-Virus 7.6 and MS
Excel.  Most of the instruments we have had for over a year and the last
two IRMSs arrived within the last year.  One computer acts as a print
server to all others.  All computers except for System 1 are DELL OPTIPLEX
that came with the IRMS.  All computers have been reformatted and
re-installed with C-D partitions where the operating system (WinXP) is kept
separate from the data (Isodat installed on D) immediately upon signing the
acceptance certificate and the engineer walking out the door.  I have
already explained this in a previous posting and the only difference now is
to install all Windows SP including SP2.  I restore the Version saved with
the Version Handler, but not the way that is recommended (I'll explain in a
future e-mail).  What kind of support do I get from Finnigan? Same as
everyone else, I go to the web site and down!
 load the service pack.  The difference is the way I report problems back
to Finnigan.  I usually write a detailed description of how the error
occurred and include the Windows Error LOG file to help the programmers
pinpoint the problem.  If they can't reproduce the error in the factory
then they can't fix it.  With the log file it helps them track it.  But I
make sure the problem is not with my system first (virus scan, loose or
broken cable etc) I have seen many bugs repaired and included in an Isodat
Service Pack, that I have submitted, sometimes within a matter of a few
days to a few weeks.

We have all the possible peripherals (except IsoLink) and some third party
software divided between 5 Deltas.  All have Isodat 2... Between the 5
systems we have lost about 2 samples so far (peak jump did not work (?))
and the systems have not crashed in the last 12 months, total of analysis:
18000.  The systems will sometimes hang up (not crash) in Workspace, when
reviewing samples on the same computer while Isodat is aquiring data, so we
try to minimize using it while running.  Re-booting every month, whether
needed or not, seems to keep the systems happy and performing.  These
numbers talk for themselves.  Below is a general view of each of our setups
without the extras that may be added to each system (e.g. special
experiments).

System 1:
- DeltaPlus (1997), Dual-inlet, Multiport, Microvol and ConFlo III, older
electronics, GPIB (ISA) communication board in a clone computer built by
myself (motherboard TYAN, 1GHz, 512Meg RAM).  I did not want to buy another
GPIB board or a brand name computer (too expensive) so I built a $400
computer 3 years ago to replace the Gateway 200MHz PII.
- TOC Analyser quantitative WinTOC software running in same computer ( 1
COM port).
- Elementar Vario EA with VarioEL elemental software running in same
computer (1 COM port).... Recently moved to System 5
- All systems have run many times simultaneously where EA work was done
while 13C isotopes of DIC-DOC was running on IRMS and vice-versa.  Also all
three ran individually dual-inlet, EA and TOC on one or two occasions.
- Type of analysis: CO2, N2, SO2

System 2
- DeltaPlus Advantage (2004) with ConFlo III and CE-1110 EA dedicated to N2
and CO2 analysis
- Type of analysis: CO2, N2

System 3
- DeltaPlus XP (8 collectors for atm work) with GasBench II, 2 traps and
sampling loop.
- Communication: 1 COM port to GC-PAL autosampler.
- Special scripts for GasBench created by us.
- Used for 18O and HD in water, dissolved O2 in water, N2O from bacterial
nitrate reduction, atmospheric O2, Ar, N2 (e.g.ice cores) and/or CO2.
- Type of analysis: O2, N2O, CO2, N2, Ar, HD

System 4 (2004)
- DeltaPlus XP and TraceGC with GC-TC (CO& HD) and an older HP6890GC(1997)
for GC-C (CO2 & N2)  including a PreCon (CH4, N2O, CO2)
- Special scripts for PreCon created by us.
- Communication: 2 com ports for GC-C and CTC A200S autosampler.  2 USB-COM
cables for GC-TC and GC-PAL
- Type of analysis: N2O, CO2, N2, CO, HD

System 5 (2005, install in progress)
- DeltaPlus XP (Special collector setup, 7 collectors), ConFlo II and
Dual-inlet
- GasBench II with traps and sampling loop, Dual-Wave laser (CO2, 213nm)
- COSTECH EA with induction furnace adapter (HTP) for Thermal-decomposition
(pyrolosis)
- Elementar EA, moved from System 1 for large samples (organic and
inorganic sulphur).
- Communication: 1 COM port for Elementar, 1 USB port for COSTECH.  Both
sofwares running in same computer as Isodat (VarioEL and EAS32)
- Type of analysis:CO2, SO2, CO, HD

AS A FINAL NOTE
The continuing development of Isodat NT is a tribute to the openness that
the programmers and administrators have had in allowing us (different
scientists worldwide) in having a say in its development and power.  Three
meetings have taken place in Washington over Isodat's development since the
year 2000 where some of us have been privy to new stuff and to add our own
concerns and ideas.  Holger Jelinsky and Andreas Hilkert have been
instrumental in allowing this to happen.  Presently Isodat could run any
instruments, from the competitor's machines to turning on your coffee maker
thanks to Holger's development of ISL (Isodat Scripting Language) and an
amazing interface for making new devices (Active screen diagrams to control
them in methods, sequences etc)  At first intimidating but extremely
powerful.  For the novice the scripts are already made, for the advanced
user with external devices to control, the possibilities are endless.  I
can reprocess data from my office while!
  Isodat is running a new set of samples through our intranet.  I have
never seen Isodat crash when I do this.  As Isodat gets even better so will
the capabilities.  Already advanced users are discussing sharing scripts on
a users website to be unveiled at a later date.

Gilles

Gilles St-Jean
Chercheur / Research Scientist
Université d'Ottawa / University of Ottawa
Sciences de la Terre / Earth Sciences
140 Louis Pasteur, Ottawa, Ontario, Canada K1N 6N5
Tel: 1-613-562-5800 xt 6830 (Bureau / Office)
                    xt 6839 (Bureau / Office Lab)
                    xt 6836 (IRMS lab)
Téléc. / Fax: 1-613-562-5192
Courriel / E-mail: [log in to unmask]
Toile / Web: www.isotopes.science.uottawa.ca