Friday, February 25, 2005

Institutional Information Technology Strategy Project

Doug,
Please post this as anonymous. I'm interested to read comments
concerning this memo....
__________

From/MS: Richard A. Marquez, A108
Phone/Fax: 7-1973 / 7-5624
Symbol: ADA-05-021
Date: February 24, 2005


Subject: Institutional Information Technology Strategy Project

I am pleased to announce that the Laboratory is moving forward
with the Institutional Information Technology (IT) Strategy
Project. As you recall, several initiatives in 2004, including the
review of General & Administrative budgets and efforts to reduce
the cost of doing business, identified the need to develop an
Institutional IT Strategy. The Executive Board reaffirmed this
need at the December strategic planning retreat, and identified
the Institutional IT strategy as one of the Laboratory's FY05
Corporate Performance Objectives. The project goal is to develop
and implement a business-driven institutional IT strategic plan
that positions IT as a strategic asset and defines the framework
for effective management of IT. This strategic plan will provide
an important context for Laboratory decisions regarding IT
investments, IT governance and organizational structure.

Project Philosophy
A fundamental philosophy behind the IT Strategy Project is that
the Laboratory's mission, goals and objectives must drive IT. This
IT-to-business alignment is paramount in viewing IT as a strategic
asset that must be managed, much like other tangible assets such
as facilities or the workforce. In support of this asset
management philosophy, the project will reinforce that knowledge
and information are enterprise assets and must be readily
available to stakeholders. Other objectives of the project include
the need to reduce the complexity of the Laboratory's IT portfolio
through reduction of the number of operating systems,
applications, etc.; and the need for optimization of acquisition
and use of IT resources through purchasing standards, acquisition
planning, efficient allocation of resources, etc. The IT Strategy
Project will address these issues and concerns in the coming
months.

Team Participation
The IT Strategy Project team is comprised of representatives from
the technical and business sides of the institution. Chief
Information Officer (acting) Charlotte Lindsey and Camilo Perez of
the Chief Financial Officer (CFO) division are leading the
project, and Principal Deputy Associate Director for
Administration Carolyn Zerkle is project champion. The team
appreciates the importance of ensuring that the Laboratory's
corporate IT Strategy incorporates and addresses the diverse needs
of the institution. Therefore, the project includes a number of
avenues for gathering and implementing the input and support of
key stakeholders.

In the coming months, managers and other personnel will be asked
by the IT Strategy Project team to participate in reviews, focus
groups, virtual forums and a variety of other mechanisms to
provide input into planning and execution of the institutional IT
strategy. I urge managers to participate and to encourage members
of their staff to participate as well. In this way, we will help
ensure that the Laboratory's IT strategy is representative of the
needs of the institution and is driven by the Laboratory's
mission, goals and objectives.

Timeline
The project timeline is ambitious. Assessment of the Laboratory's
current IT environment already has begun with the assistance of
Gartner, Inc. and Decision Applications (D) division. In the
coming weeks, the team will develop a process for organizational
assessments of IT resources, at which time managers, or their
designees, will have opportunities to participate in the project.
The Laboratory plans to begin implementing a corporate IT strategy
this summer.

Quick Wins
The team also is developing a set of "Quick Wins" institutional
IT improvements that can be accomplished in the next six months.
The quick wins include:

- development of an IT cost tracking system,
- implementation of a desktop computing purchasing and
lifecycle management system, and
- implementation of a tracking system for software licenses
and license purchasing.

Communication
To keep the institution updated on the status of this project, the
IT Strategy Project team has developed a website located at
http://int.lanl.gov/cio/itstrategy/index.shtml that includes
frequently asked questions about the project and other resources,
including a planned moderated discussion forum. Those with
questions about the project can contact Project Leader Camilo
Perez at 5-8704 or submit questions and comments to
itsp-faq@lanl.gov online.

Comments:
I have read this twice and still have no ideas what the heck it means. IT must be a new word learned yesterday by the incompetent author. Idiot
 
Sounds scary - The managers on the IT Team don't have a clue about IT and having Gartner invloved is what clueless managers buy in to. The "Quick Win" items are already being done more or less. So, We will need to see if the rest bears fruit. We are going down the toilet real fast.
 
It probably means you will only be able to use WindowsXP and
mostly Microsoft products to accomplish the work on your
desktop. If you need to use highly specialized products
from niche software companies, then fuggetabaoutit. If you
find it useful to use shareware or various open-source
products to accomplish your work, then fuggetaboutit. If
you enjoy using truly secure OS's like Linux or MacOS X,
then fuggetaboutit. And if you want to deviate from the
"approved" list of software products, then that software
will first need to be "tested" by your computer security
officer, and then, perhaps, you'll be able to use it. Of
course, this testing will take lots and lots of time. You
have lots of time, don't you?

I've seen policies like this put in place at other
institutions. It ends up destroying the ability of staff to
get much of their work accomplished. If you work at GEICO
Insurance or perhaps at Acme Widgets, then it works OK. If
you work at any place that does unique data analysis or
scientific work, it usually ends up stopping you dead in
your tracks.

A while back, Nanos gave a year-end "pep-rally" talk. You
remember the one with the cheesy movie clips from movies
like Rudy, et al. During that talk he mentioned the need to
get the staff off unknown products, and that we should
standardize on "enterprise class products like those from
Microsoft". Our brillant Director wants us to standarize on
products that are buggy and known to be a source of attacks
from almost every hacker on the planet. Well, his wish may
soon be your command. If you find that you can't do your
work using only MS Excel, the fuggetaboutit. Besides, MS
Office is all the bean-counters really need to accomplish
their work, so it should be good enough for you people in
the scientific staff.

I hope I'm wrong about the above, but our history over the
last year will probably prove that this is exactly the
bone-headed direction we are moving to with this newly
announced "IT strategy". It basically means more control
from clueless people up above telling you how you will
accomplish your scientific work. Better get use to it.
 
It's a power grab from IT, driven I am sure by Pete and his "enterprise class apps from Microsoft" ignorance. The navy loves microsoft. I expect to have to deal with the CIO interference very soon now.

The CIO claims to have technical people on board but no clueful computer people seem to be on this comittee. I think they've reached their conclusions already and this process is a rubber stamp.

If it gets too bad, well, I do have standing offers at other DOE labs and several corporations, so I've got options. Of course, I'm not the only one. I'm giving this place one more year and them I'm done.

Pete is beginning to serve as a mediocrity filter. The Lab is getting to be a tough place to do good work, due to initiatives like this one. I've seen a lot of top guys leave in the last 9 months -- yes, it's only 10 people, but 10 LEADERS -- the kind of guys you can't afford to lose.

I have seen this at other dying institutions. Drive out all the people you don't agree with and you have a nice collection of yes-men and hacks at the end. I wonder who gets to turn the lights out?
 
This project is contrary to the well-being of the country and the success of many LANL projects if it is applied to scientific computing. The description of it does not support the objective of having LANL's mission, goals, and objectives drive the IT project. In fact this project will drive LANL's mission, goals, and objectives right into the ground unless there is an understanding that it will be applied only to those desktops used by admins, not technical staff.

Who are the members of the IT Project Strategy Team? What are their backgrounds? How many have done scientific computing at LANL (as opposed to powerpoint presentations on Microsoft products? Do any of them realize the complexity of validating certain codes on new hardware/software? Do any of them realize that standardizing on something today is a colossal waste of money because tomorrow that something will no longer be supported? Do any of them realize that those doing the work are the ones best able to select the tools with which to do it?

Deciding before the project even starts that an objective is to reduce the number of operating systems, applications, etc. is putting the cart before the horse. It clearly shows a lack of understanding of how much in flux the entire computer industry is. It can not possibly work unless someone wants to send LANL millions of dollars every 3 years to completely upgrade every computer on site.

This is a formal endorsement of an IA project which was also poorly thought out but is unenforceable. There has been no acknowledgement that, as computers age, newer hardware does not function nearly as well if at all, so before operating systems can be replaced, hardware has to be replaced. New hardware purchases drive the use of certain operating systems. This is how the industry as a whole forces the replacement of operating systems, and for the most part it works quite well.

This project will also contribute to the death knell of northern New Mexico business by developing "purchasing standards, acquisition planning, efficient allocation of resources, etc." which means more purchasing from a monopoly rather than using local vendors.

LANL has been on the forefront of scientific computing for over 45 years. It has never needed a business project to force a move from one hardware/OS to the next. Ten years ago there were many NeXt computers in offices. Today they are gone. There used to be many Hewelett-Packard computers. Now most of those are gone. SGI's running IRIX are slowly making their way to salvage also. There are no longer any Sun Sparc stations running SunOS; older Sun hardware is gone; and the newer hardware runs later verions of Solaris. As the market changes, what's on the desktop changes.

Finally, there will be a number of projects requiring that which is not covered by this project. Using them will require inches of paperwork for exceptions. This will be another huge waste of taxpayers' dollars in the same vein as the Enterprise Project.

ADA Marquez would make better use of his time if he ensured that the business side of LANL worked more efficiently, paid its bills on time, not weeks late, issued PR's quickly, etc. In fact, ADA Marquez should investigate all puchasing at LANL with the goal of doing as much as possible locally rather than with monopolies. And he should investigate the harm being caused to northern New Mexico businesses by purchasing decisions such as the one that forces all contracts under $500,000 to KSL rather than local companies.
 
This whole "IIT Strategy Project" is just another extension of Nanos' inability to provide appropriate leadership at LANL. Pretty much everything about it is just plain wrong, but until Nanos is removed, Marquez has his approval to try to shove it down our throats.

I suggest passive resistance until we get better management at LANL.
 
I love the first comment by the guy who feels their opinion has value and doesn't even know what the term "IT" means. Classic. I don't understand it so the other guy must be an idiot.
 
The Lab should specify right away that this IT initiative should be for business employees only and leave the scientific computing to the scientists and engineers where it belongs. Turning the Lab into a Dilbert environment is ludicrous.
 
Yeah, let's preserve lassaiz faire in order to protect "scientific computing."

This is the same scientific computing trying to keep an ancient, wheezing Windows 3.1 machine supported because it's running the original software necessary to the research effort, and the researcher hasn't bothered to explore any upgrades over the years.

This is the same scientific computing that thought hand-carrying classified data to work on at home was OK before WHL.

This is the same scientific computing that is still coding in FORTRAN because the fossils that first wrote the code are entrenched and can't be bothered to learn a modern language, or allow their code to be reverse engineered. It doesn't matter that every young graduate coming into the instutution is versed in C/C++ and has to be expensively retrained to speak in the ancient tongue. Let's stick with the old ways. We don't need modeling tools, code generation, automated testing framework generation, test robots or any of that other non-scientific stuff that these young upstarts have gone soft on.

This is the same scientific computing that feels they are qualified to determine what plant engineering and automation software is capable of supporting enterprise-class operations, despite having no actual plant operations experience.

This is the same scientific computing that believes using a piece of software from sourceforge.net is OK because it's open source, but don't bother to check the fact that the project has only two active participants, is still in alpha, and hasn't been revved in over six months. Yeah, that's quality stuff.

This is the same scientific computing that feels they are qualified to determine what document management software is capable of supporting enterprise-class operations, despite having no actual content management experience.

This is the same scientific computing that can't pass an audit to save their life.

This is the same scientific computing that considers quality assurance to be an "unfunded mandate."

This is the same scientific computing that thinks network security is a trivial matter, despite more than one hack attempt per second, 24/7, on our firewall.

This is the same scientific computing that doesn't think about foreign intelligence networks having a map of every IP address serving the Los Alamos area outside the firewall, and still believe their home computers are an appropriate place to store and work with corporate information.

This is the same scientific computing that sets up a LINUX workstation with all the security features disabled to make it more convenient for team use.

This is the same scientific computing that thinks it's OK to have 25 different issue management software products at the Laboratory.

This is the same scientific computing that thinks it's OK to buy a new computer and move their code over without testing it.

This is the same scientific computing that can't even show so much as a test plan for their code.

This is the same scientific computing that demands security and infrastructure support for at least four main operating systems, yet complains when overhead costs for computing infrastructure are higher than the industry average by a factor of at least two.

This is the same scientific computing that can debate the differences between statistical software packages all day long, but can't understand how modern concepts like XML are obviating the argument.

Oh yes, let's not try to behave like every other professionally managed organization so that "scientific computing" won't be endangered. Puhleeze.
 
The previous anonymous author is absolutely spot on.

As I told one of out scientific types ....

" I don't tell you about particle physics, you don't tell me about computer security. "

He got the message, and we have a good working relationship.
 
A previous poster said: "Sounds scary - The managers on the IT Team don't have a clue about IT and having Gartner invloved is what clueless managers buy in to."

The managers on the IT team have massive clues about IT. They live and breath this stuff. Most have Masters degrees in the field. Most have come up from the grunt level, instead of stepping in as a TSM right out of college. They know what's right and they know what's wrong. Gartner is a research and facilitation contractor so they don't have to spend time re-inventing the wheel.

The problem is funding, as usual. These kind of changes will cost massive megabucks, and if improperly planned, can cost as much as an order of magnitude more before the failure is evident. When done successfully, the literature overwhelmingly shows massive cost savings over the former operational environment.
 
The ADA office is composed of a bunch of idiots who have no leadership skills whatsoever. Marquez, by the way, is a known harrasser and misogynist from DOE Albuquerque who of course was hired by Los Alamos. He couldn't even define IT if he tried. The CIO is a nice person but is incapable of making decisions and won't stand up for herself or her people. Wasn't the Lab IT plan supposed to be completed by the end of 2004, but instead they've been spinning their wheels on this for nearly a year? The only competent person -- who possessed a Harvard MBA [and was quite a cook, I hear] -- in the CIO Office retired and left Los Alamos in January in protest due to racism in that office against him as he was native american, and because his wife -- also an award-winning Ph.D. scientist -- was forced to resign from CCS Div. for speaking her mind and criticising her managers. Marquez and Co. are the same people that are so-called in charge of the EP project that is behind by 3 years and has more than doubled in cost. This sure is a cluster but not a linux cluster...
 
I wonder who is the smug guy with the long list of identically-formatted paragraphs of straw man arguments. He is pretty good at stringing together sequences of buzzwords in a way that is simultaneously impressive to amateurs and vicariously embarrassing to experts. As someone who works with XML daily, I see the claim that XML "obviates" the need for appropriate statistical software as laughable. XML is a data format, like HTML. It won't cook your soup or calculate the mass of a neutrino or any such thing. How long before someone claims we can replace Linux with XML?

I see he also has a shadow, who posts a me-too response each time, as if that were clever.
 
There is only one "IT plan" that needs to be accomplished. That is to get
a RedNet computer into the office of every Q-cleared person that works
behind the fence. On this one, the Lab is failing miserably. The cost,
paperwork, and time it currently takes to get a RedNet port put into your
office is completely unacceptable. It took me around 8 months, paperwork
from about a dozen "support" people (what a joke), and around $18 K to get
a port into my office. And what of the $20 million that DOE just gave
LANL to begin spreading out the Red network? It seems that most of this
funding is going into glossy RedNet posters touting this network that I've
recently seen sprouting around the Lab walls. I hear the money was transferred
to upper level Program Manager types. I'll bet the $20 million gets burned
doing some useless white paper studies. It should be used to start pushing
out more RedNet hookups ASAP.

If there was one single Lesson Learned from the whole CREM incident, it
was that we need to make access to the RedNet easy and available to the
whole Q-cleared staff. On this one, IT management is failing us big time.

In addition to the RedNet, it's time our staff joined the rest of the
US government and hooked more classified systems into the rapidly growing
DOD SIPRNET, so we can efficiently contact people throughout the classified
world. Many other agencies, such as the FBI, are jumping onto the DOD's
SIPRNET bandwagon as a safe, efficient means of classified communications.
Why not LANL? Who are the gate-keepers who are holding us back on these new
communication innovations? What do they fear they lose by moving forward?

No, I guess these things are too hard for IT management to handle. They
would rather make sure we all have our desktop PCs configured to their specs.

We need to start by fixing the big problems first.
 
So far this blog has demonstrated a remarkable self-regulating behavior. If somebody posts something stupid, or just way out in left field, a bunch of contributers jump on it by way of correction, and all without flames.

Admirable (in the good sense of the word, not the Nano-sense).
 
Agreed on RedNet. The Executive Board has a short attention span for essential activities, and too much time invested in pet projects they read about in trade magazines. These people are too well paid to be wasting their time pontificating on Fortran or Linux.
 
What are the security implications of pushing Microsoft software at the Lab given that over 2000 Microsoft software developers are outsourced to India, a nation on the DOE Sensitive Countries List?
 
"Information Week" magazine had a great lead article just last week about
the continuing problems with serious bugs in the Windows OS.

.....................................................................
You Call This Trustworthy Computing? - Feb. 14, 2005
Three years into Microsoft's security initiative, the bugs keep coming

Microsoft last week issued a dozen security bulletins addressing 17 software
vulnerabilities, tantamount to a shotgun blast of holes through the company's
product line. Nine bulletins, many graded "critical" in importance, affect
various versions of Windows. Others address problems with Microsoft's .Net
Framework, SharePoint Services, Windows Media Player, MSN Messenger, Internet
Explorer, and Office suite.

Even Microsoft's most-secure operating system, Windows XP Service Pack 2,
wasn't immune: More than half the bulletins involve SP2. To repair all the
vulnerabilities in all affected products would require more than 60 patches
on English-language computers alone. "It's an almost endless list," says
Kyle Ohme, director of IT with Freeze.com, a Web-site operator that uses
about four dozen Windows servers, some of which are IBM blade servers,
to offer screen savers to millions of users each day.
.......................................................................

Yeah, LANL IT management. Let's enforce a single OS. And let's make
that OS a very buggy, hacker-prone OS. That's the ticket. That way
our IT staff can be permanently employed over the next decade making
patch after patch after patch just to keep our systems marginally safe.
Sounds like a winner to me.
 
Does anybody remember Inanos' remarks, "every Unix program has back doors, but Windows is 60% tested"? We're really on a ship of fools if this silliness has started to become official policy. Maybe the Director will issue a document indicating which 60% is "tested."

It's one thing to be ignorant and curious. It's quite another to be ignorant and stubborn.
 
Anyone, including Microsoft, who believes he or she understands 60% of Windows XP has never looked at the source code for that program which has more lines of code than we have written here for the weapons program during the past 60+ years. BTW, who would have the gall to brag about having locks on sixty percent of his or her doors? When was he last time a virus hit Linux and Apple systems?
 
There have been a number of derogotary comments made on 'the' blog about ancient languages like Fortran. Can some one comment on the truth to the story that I heard about a code group at the lab ( I don't remember if it was an X or CIC or CCS or CCN or ASCI effort ) that tried unsuccesfully to translate some of the fortran code over to C/C++ and they could not get it to run as fast as in fortran. Some staff claimed that it was due to inherent problems with the C compiler and others said it was just inept fortran programmers trying to speak a secong language.

Just wondering.
 
For most scientific computing, involving iterations or equation solution, C++ cannot produce as efficient an executable as Fortran. Each language has its own domain of applications where it should be used.
 
I know for a fact that the Red Network expansion was presented to upper management, including Vann Bynum, more than two years ago.

I know for a fact the security benefits were clearly and succinctly promoted, including reduced or eliminated CREM and the attendant risks.

And I know for a fact that it was substantially ignored.

Until CREM-III finally rose up and made it the hottest thing in the place, and why-didn't-we-think-of-this-sooner.

Some leadership.

Now they want to lead us to the future of standardization with Gartner Group?

In my 22+ years of work in information systems, I've never seen Gartner be important for anything but pithy quotes in trade rags.

To me, they're the Paris Hilton of IT...famous for being famous, but seldom if ever demonstrating any real talent for anything.

A perfect choice to match our leaders.

Dilbertization is under way.
 
Someone said: "For most scientific computing, involving iterations or equation solution, C++ cannot produce as efficient an executable as Fortran. Each language has its own domain of applications where it should be used."

Fossils have been doing the efficiency dance as each ancient tongue has gone by the wayside. Machine language is faster than Fortran. So is Assembler. Why not use them in the name of efficiency? Because you don't understand that particular language, and for good reason.

Now, I'll admit it doesn't make sense to wait for Moore's Law to compensate for the use of an "inefficient" language, but has anyone there heard about component development? A simple analysis points out where the iterative work is done; optimize that as a separate component. Let the rest of the team feed it using components built using modern languages. You might actually get something worthwhile out the door in a reasonable amount of time.

But maybe that's not the objective.
 
I have an idea: Let's all use LINUX! That way we'll be dependent on those few vendors that actually have produced useable stuff for that platform. Typically, if the main vendor goes south, the alternatives are few and far between, but the main vendor never goes south, so that won't ever be a problem.

And if what you need isn't written for LINUX, or there's only one really sucky package out there? Well fuggedaboudit, you don't really need that stuff anyway. It's not real science.

It's interesting that the only times I've seen a scientist speak in favor of computing security is when they they might be forced to use Windows.

Let's get the argument down to two platforms, shall we? Windows has won, hands down, for the administrative / operational side. Too much in place, too many people trained to go elsewhere now. So why don't we shift the debate to deciding the standard platform for scientific computing? Those incompetent IT weenies can stay out of it. I'm sure you genius scientists can arrive at a well-founded solution very quickly.
 
To the person who said (relating to Fortran and C++) " Let the rest of the team feed it using components built using modern languages" - that's exactly the problem! You need a team of C++ programmers to do what a good Fortran programmer can do alone, and in much less time. This is an example of the "unbridled arrogance" that provides work for a team of programmers replacing one person who thinks about the problem. People are entrenched in what they are doing, and obtuse to analyzing and employing the best, even if it is not fashionable in some circles.

However, this is not the right forum for this. Let's concentrate on getting LANL going!
 
"It's interesting that the only times I've seen a scientist speak in favor of computing security is when they might be forced to use Windows."

Perhaps that is because many scientists know just how bad the security holes in Windows really are.
The main problem with Windows is that far too many "features" are embedded too close to the OS, such as with IE and lately, with Windows Media Player. BTW, have you applied the latest 60 patches to your WindowsXP SP2 system yet?
 
There are plenty of Macs being used effectively on the operational/administrative side at LANL, and they require less maintenance and support because 1) there are no viruses or worms in the wild for the platform, 2) the hardware base is uniform, and 3) Mac OS X has better "fit and finish" than Windows.

Operational and security concerns would militate for transitioning off the Windows platform whereever possible. How much time and money do we spend dealing with the daily batch of new Windows worms and viruses? And how easy is it to get administrative users to click on some windows script, .exe attachment, or ActiveX control that will then come right on in past the firewall? How could this time and effort be put to better use if everybody used Mac OS or Linux?

"Windows" and "security" are two words that cannot be used in the same sentence respectably.

-dug
 
"However, this is not the right forum for this. Let's concentrate on getting LANL going!"

I agree that the assumed point of this blog is to feel our way towards a better day at LANL. But the displayed polarization over topics such as operating systems for computers is a symptom of our bigger problems. 'We' feel like 'we' know what is best for 'us' in 'our' research efforts. Therefore 'we' know what is best for everyone else as well.

I really laugh when some well intentioned staff suggests with honest zeal that I should stop programming in a language that I have worked in for 20+ years so I can be 'modern'. If one of the big code projects wants to be a mix of C, Fortran and what ever else, then fine. If there is a reason for it. Lets be better at doing things that are defensible based on facts not our favorite belief system.

But if I am working on a multi-lab, DoD/DOE/academic collaboration that is based on a huge piece of legacy code ( read that as Fortran ) then we will not overly complicate the process by rewriting large portions of the code just to be 'modern'. We have a deliverable, a set budget, a set delivery schedule, etc. Adding an unfunded mandate that the customer doesn't want does not make sense nor is it justifiable.

But, I should not be suggestiong that I know what is best for every other researcher, operations person or support staff at the lab. As professionals, we should each be doing what is best for our areas of expertise.

I guess that might be what rubs some people wrong. Someone who 'seems' to have never bothered to ask any questions about what I do gets to tell me how I should be doing it. I think that is why the 'mushroom' analogy resonates for so many lab employees. We are kept in the 'dark' ( i.e. relatively uninformed by management), we get fertilized ( directives, mandates, revamped security and safety paperwork ), we try to grow ( deal with the 'crap' as best we can while getting our work done ).
 
"Lets be better at doing things that are defensible based on facts not our favorite belief system."

"As professionals, we should each be doing what is best for our areas of expertise."

Now we're getting somewhere. So who would be most appropriate for setting IT strategy?
 
LANL's computer arrogance has permeated even the admin side for years and years. I dated a coder in 1985 who worked in the admin area in the Otowi building. She was writing an application from the ground up to do LANL travel in COBOL!

The same absurd mentality is killing the Enterprise Project today. LANL is so "Special" that they can't use canned, proven, broadly-installed Oracle applications. "We are SO SPECIAL that commercial software applications for non-scientific uses are rejected in favor of custom, buggy kludges. It's nice to play in a cost-plus environment - just ask the vendors for the IRS and FBI systems.
-Dawn-
 
"Now we're getting somewhere. So who would be most appropriate for setting IT strategy?"

You beg the question. Who says we need a single, ideologically-driven IT "strategy"? How is some IT "expert" on the mountain supposed to know what is appropriate for the particular circumstances of my project? Support functions at the Lab are supposed to do that: support, not carve off little empires for themselves, "We *own* computing," "We *own* decision making," "We *own* project management."

What's next, setting up a Lab division of clothing experts to decide Vestment Strategy?
 
Enterprise systems were a fad that peaked around 1998. Many large businesses have been backing out their failed ERP systems and returning to their legacy systems, at great cost and lost opportunity.

"Enterprise" is not something you can buy in a box. Every organization is different, and this forces the ERP vendor to add lots of one-off kludges to their software. Sometimes Company A's kludge conflicts with Company B's kludge and you're not sure who's going to win in the next version. This tends to interfere with the promised benefit of getting "upgrades for free."

ERP systems are generally designed for commercial entities, not matrixed non-profits. Sandia had to hack their system quite a bit to make this work.

For 5% of what EP has cost LANL already, we could have contracted out a custom solution and had it implemented already. Business managers not familiar with the fine-grained details of their organization tend to be wowwed by the slick salesman and glossy brochures, and then hector their internal experts as change-resistant when they try to inject some reality into the decision. That is why ERP disasters have played out through a great many of the Fortune 500 companies.
 
From the length of this blog's comments, it's obvious that scientist are
passionate about their computers! And what's wrong with that? After all,
it's the main tool used by a large portion of the scientific staff.

I would not worry about this particular area of the blog veering off the
main topic. I'm finding the comments in this subject area to be a lot of
fun to read. People will defend their favorite OS and programming language
almost like it was a religion. No problem with that. I used to program in
Fortran, but now use C/C++. However, there is no shame in using Fortran.
Many of the best ideas in a particular language usually cross-fertilize
into the other languages with time.

Most of the scientific staff undoubtable feel that IT support should NOT be
enforcing their desktop and application decisions on us. I have to use
Windows, Linux, and MacOS X to do my work, and wouldn't want to lose
any of the three. Windows has the lion share of apps (but can be prone
to many problems), Linux has a huge base of very good open-source apps,
and MacOS X gives you the best of Unix, a moderate app base, and an
extremely polished interface.

Mono-cultures are a bad idea in computers, as in biology. Don't take
my Linux/MacOS away from me, IT staff. Them's fight'n words!
 
The ITSP is now the lead masthead story over at the LANL News Bulletin, so you know it's important with top LANL management. It says that it will reduce the number of OS and applications in use at the lab and effect virtually EVERY Los Alamos employee. I don't think this is just for the support staff. If you're a scientist here and you enjoy using Linux or MacOS X, then watch out. Nanos and company may be coming to shut down your computing options.
 
Hopefully this just means that they'll try to get rid of Windows 95 and 98 and Mac OS 9 and the like. There are huge numbers of Mac and Linux users (administrative and technical) at the Lab. The entire organization is inflamed enough without them trying to ram some ill-conceived program down everybody's throat. If they try to push something like all-Windows, all the time, this place will just crack wide open.
 
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?