Friday, May 6, 2005 (A.G): OpenOffice Can Save Documents As PDFs

On April 12th, put out by Adobe's plans to acquire Macromedia, I was musing about how nice it would be to have some open-source alternatives to Adobe Acrobat -- which I contend is the most flakey piece of allegedly production quality software that I have had the misfortune to use in 36 years in this game.  So it was serendipitous to hear that the FREE, OpenOffice suite (spawned off by Sun) allows you to save documents as PDFs.  So I had a quick look at OpenOffice, which I had not looked at seriously in a long time.  I liked what I saw (and read).

Yes, I realize that it currently does not have an Outlook equivalent and that is a fairly major impediment when it comes to enterprise use given its wide use for non e-mail applications [e.g. scheduling]. But if you can get around that this suite might definitely be worth looking at, especially since it is a bona fide open source project.  I don't have to go on about the 'joys' of Office.  We all know the score there.  This really could be a genuine option for many of us.  I was intrigued that it includes a vector drawing program, i.e. Draw, in addition to Impress, the PowerPoint equivalent.  There is even a mathematical equation generator.  So lets all download this and try it out.  I will be honored to publicize your feedback on these pages.  Have a great weekend.

Thursday, May 5, 2005 (A.G): 13,000 Layoffs At IBM Just Isn't Enough!

Job losses, in any sector, invariably grieves me.  But I have to say that I was overjoyed to hear that IBM intends to shed 13,000 jobs – especially since it came just 2 days after I had shared with YOU my growing concern about IBM and Palmisano.  IBM is bloated again with a surfeit of dead wood.  13,000 isn’t nearly enough particularly if that will, at most, only result in $500M in long-term savings.

10% of IBM’s now 300,000 is incontrovertibly crème de la crème.  It is thanks to these folks that we see things like the POWER5.  20%, in my reckoning, try hard and have an acceptable work ethic.  10% is counting the hours to retirement while another 10% are not sure what they are supposed to be doing at IBM.  But that leaves us with the remaining 50%.  Now remember I am an ex-IBMer and have also had lots of dealings with these folks.  Thus, what I am now going to say is not easy and not something I am doing lightly.  But it has to be said.  This 50% is underproductive and lazy!

They continually claim to be so busy but produce nothing.  All excuses, never action.

This is the bane of today’s corporate world.  People that just don’t get it that being busy does not equate to being productive.  IBM’s true productivity sucks.  For a start they should get rid of 75% of their so called marketing folks.  Let the products, many of them exceptional, speak for themselves.  These folks do more harm than good.  This will save more money while increasing revenues.

Things are not good at IBM, overall.  Palmisano has let things slide.  The dancing elephant that Gerstner left behind has become flat footed and clumsy.  So lets not stop at 13,000.  Make that 31,000 and that still will not be enough.  Sorry.  But somebody has to say this.

Wednesday, May 4, 2005 (A.G): The Next "BIG Thing" In Big Iron IT

Enterprise Workload Management (eWLM) is perspicacious, technically rather sexy (albeit in a geeky sort of way), but above all a groundbreaking technology to solve a genuine problem.  eWLM recognizes that today’s Web-centric transactions are invariably going to rely on multiple, most likely disparate servers – particularly so if a mainframe is involved.

Trying to assiduously manage transaction performance on one platform, as we try to do on mainframes with WLM and even IRD, is futile if the other servers involved in the end-to-end delivery of a transaction are totally oblivious of the business priorities in play.  What is the point of a mainframe LPAR giving CICS transactions priority over DB2 queries, if the ‘front-end’ portal server responsible for end-user delivery does not know that CICS transactions are the most mission-critical during 9-to-5 – East Coast time?

Once you start thinking about the Web servers, portal servers and application servers that are likely to be involved in the processing of a typical big iron transaction you can immediately start to see what eWLM is trying to deal with.  The rest kind of falls into place.  This is why I am becoming increasingly besotted with eWLM though I know that what we have today is barely embryonic.

But this is a good time to get acquainted with it because the technology curve, as yet, isn’t that steep and easy to master.  eWLM is featured in the new WebSphere Application Server for z/OS V6.0.1.  But all we have right now is essentially reporting and monitoring of transaction performance, across conformant platforms.  So today it is a glorified but incisive adjunct to response time monitoring.  This will, however, start to change next year with the availability of z/OS 1.7 and the next generation of eWLM ‘managers’ on other IBM platforms.  Down the road eWLM will even help us load balance across different instances of servers.  So there will, starting next year, be rapid strides being made on this technology.  Trust me on this.  Get to know eWLM now.  It will serve you well down the road.

Tuesday, May 3, 2005 (A.G): Disquieting Discord from IBM

We have all heard, incessantly, over the last couple of weeks, vis-à-vis JP II, how hard an act he would be to follow – particularly for mere mortals.  It reminded me of another very difficult succession of late, and no I was not thinking about Welch.  Well, when they also started mentioning miracles, I, a devout (and professional) IBM watcher for 25 years, immediately started to feel very nostalgic about Gerstner.  Now that was a class act (despite the recent accusations that he microwaved some of his numbers) with a string of documented miracles.

Palmisano, though a sax player, was thus, at best, always shaping up to play but second fiddle.

Of late, even that second fiddle hasn’t sounded right.  When you have been at it as long as I have you pick up on the nuances, the change in timbre and the low frequency vibes.  They are disquieting.

This ruckus about the firing of an army reservist by IBM epitomizes why I feel that IBM, yet again , is unraveling.  Somehow that cohesion that Gerstner someway managed to instill is wearing away.

The 1Q results that we have already spoken about (April 15 BLOG) are but the tip of the iceberg.  We also had that inexcusable charade about mainframes being idle 40% of the time and Unix servers only serving for 10% of their time.  There are other more insidious instances that I am currently not at liberty to talk about.  Lets just say that things could be much, much better.  Enough said for today.

Just remember, that I started bitching about the inadequacy of IBM’s network hardware division, in print, 2 years BEFORE IBM got around to giving it away to Cisco.  Alas, I am getting the same types of bad vibes again … but now from the server groups.  Lets hope that by some miracle Palmisano is able to fix all of this before it becomes too late.

Monday, May 2, 2005 (A.G): SOA Is Not Contingent On Web Services

The IBM networking sector, which includes host access/Web-to-host, has always been very partial to what I would like to refer to as Hare Krishna marketing – though one would think that by now they would have cottoned on that this style of marketing, more often than not, has proved to be spectacularly unsuccessful!  So what do I mean by Hare Krishna marketing (especially for those of you too young to remember HK devotees at airports or marching through the busiest thoroughfares)?  It means, mindlessly chanting the same supposedly mystical mantra over and over again, accompanied by some inane hand waving to add drama, in the hope of gaining salvation.

Well a decade ago we had “ATM”.  It used to go: “Ohm ATM, Ohm ATM, Ohm ATM”.

Then not that long ago it turned to: “Ohm XML, Ohm XML, Ohm XML”.  Well when that didn’t work as well as they had prayed, they changed it to “Web services” – sometime sticking XML in front of it, hoping I guess to cover more fronts.  How can I get this across delicately, without offending too many people.  Web services also haven’t really lived up to their much vaunted expectations.

So we now have: “Ohm SOA, Ohm SOA, Ohm SOA”.  You get the drift.

And as with Web services and XML, we have again reached back for some backward compatibility.  People who go around chanting SOA invariably try and claim that it is magically intertwined with XML Web services.  Not so.  Enough said.  If you want to learn the facts read a few paragraphs on page 11 of this document I wrote earlier this year.

Friday, April 29, 2005: Entry Was Canned!

Thursday, April 28, 2005 (A.G.):
Who Would Balk From The Philanthropic Grid?

I had an e-mail from IBM (no doubt sent to a huge mailing list of all of its contacts) over the weekend asking me to join the World Community Grid [www.worldcommunitygrid.org].  I had heard of ‘WCG’, earlier this year, at Globus – though I have since found out that it was launched in November 2004 by IBM and a bunch of the world's leading science, education and philanthropic organizations.

WCG can be thought of as SETI@home [setiathome.ssl.berkeley.edu] extended to cover a whole gamut of important issues – rather than just seeing whether ET is trying to call home.  Per IBM, WCG would use our unused computing resources (my so called PCgate scandal of IT) to crunch scientific applications dealing with issues such as AIDS, Alzheimer's, Tsunami forecasting etc.  You can insert your favorite “charity” here.

Given that I am extremely committed to grid computing you would have thought I would have jumped at this and signed up all my PCs – especially since I had no real problems letting SETI draw mesmerizing screen savers on my backup PC.

Well I didn’t, and that surprised me.  I felt squeamish!  That was a revelation.

I always knew that this was a potential stumbling block with grid computing.  In today’s security-to-the-forefront culture do you really want unknown 3rd parties running software on your machine – even if their credentials appear to be immaculate.

And funnily (or is it sadly) enough my problem was with IBM.  Now as an ex-IBMer I kind of trust IBM – implicitly.  They are not Microsoft.  But I also know IBM too well.  And that is the problem.  They might not be malicious but they are no longer diligent like they used to be.  They can be very careless.  So I let this opportunity go by.  What would you have done?  What would Gerstner do?

Wednesday, April 27, 2005 (A.G.):
Why Talk is Cheap, per Microsoft and MS/SAP “Mendocino”

This BLOG was going to be devoted entirely to “Mendocino”, the SAP and Microsoft collaboration, but another news story from Microsoft, had me lost for words – literally!  Richard F. Rashid, supposedly Bill’s top researcher, told AP that: "In English, at least, it's faster to type than it is to talk."

This was new (but immediately dear) to me, and I am more than willing to give Rashid the benefit of the doubt. Wow.  You need to think about this.  It was definitely counter-intuitive to me.

Just did a crude search on Google.  It appears that the average typing speed is 40 wpm.  .6 of a word per second.  That is funny.  At least in this country, the accepted, rule-of-thumb basis for measuring a second, using words is, of course: “Mississippi 1” etc.  Hang on.  Isn’t that two words, and one of them rather long to begin with?  I will, as time permits look, into this.

This doesn’t help me.  Though I spend around 10 hours a day on a keyboard, and have been known (when deadlines loom) to produce upwards of 7,000 ‘camera-ready’ words a day (which is 3 times more than what professional journalist are typically expected to crank-out) – I don’t touch type.  Never have.  Never will.  I peck around using two fingers.  BUT, to be fair to Rashid, I do prefer to type.  Can’t write anymore.  Have difficultly even writing a check (which might be Freudian too).  My hand muscles are no longer attuned to grasping writing implements.  I am, as a chronic, incurable, Type A++++, also accused of talking fast.  Rashid has managed to put a HUGE bee in my bonnet.  One stat. that I do know.  English verbiage is 50% redundant.  Now if we can factor these two together we might be able to come up with something truly useful for the future.

But now to “Mendocino” (and the choice of this name intrigues me given its San Francisco, as opposed to Seattle, area connotations).  “Mendocino” will link certain SAP ERP related application functionality with MS Office (and I assume that this will be limited to Office 2004, obviously as an incentive for enterprises to finally upgrade).  In particular, I gather, that you will have direct, seamless(?) access from Excel, Outlook and Word (and most likely PowerPoint and Access) to SAP functionality pertaining to budgets, time management, expense control and organizational processes.  This is good and reminds me of some of the similar integration available between legacy applications and Office.  My only concern is that this is a ‘point solution’; i.e. specific to SAP.  It would be nice if it was more generic.  But this is a good start.

Just one other point.  I did some research as to why MS/SAP would pick “Mendocino” as the code name for this product.  I, to my intense amusement, came across a company called Mendocino Software, in Fremont, CA.  Guess what they specialize in?

Recovery Management solutions!

Maybe Microsoft is trying to tell us something.  Laugh.

Tuesday, April 26, 2005 (A.G.):
NetManage Results Leave Us In The Dark

Given last week’s auspicious pairings of WRQ/Attachmate and SEAGULL/SofTouch, there was particular poignancy to NetManage’s 1Q results, that were announced after the bell yesterday.  Well, I won’t keep you in suspenders if you haven’t seen them already.  They didn’t tell us much about overall market conditions.  But that is not far from par for this course.

WRQ and Attachmate are private and SEAGULL is not on NASDAQ.  Jacada, once the standard bearer for the potential of Java, is on NASDAQ, but it is now (with revenues in the $20M range) a relatively small company and based out of Israel to boot.  So that leaves us with NETM – which despite its Israeli connections is, however, based out of Cupertino.

Given its current run-rate of around $48M (down from $66M in 2002 and $18xM in 1999 when it acquired Wall Data) NETM, has to be #4 among the traditional host access/Web-to-host vendors behind IBM, WRQ and Attachmate.  Thus, in theory, its numbers should give us some perspective as to market conditions.  In practice they rarely do – and yesterday’s numbers were no exception.

The $12.1M revenues for 1Q2005 are $200,000 more than what they made in 1Q2004.

That was suitably hyped and from what I can see most people bought that.

But my memory (even in my old age) is still better than most in this sector.  SMILE.

Didn’t NETM acquire Librados, supposedly an iWay and Attunity slayer, in September last year?

Hence, the 1Q numbers being quoted are not even close to apples-to-pears.

When you factor Librados in … the $12.1M looks anemic.  Even a bit scary.  iWay claims to be doing around $50M and Attunity though tiny states that its revenues increased 13% from 4Q2004 to 1Q2005.  Hhmmm.

Just for the record NETM’s 4Q2004 revenues were $13.1M.  Enough said.

You must get the drift.  These numbers were not indicative or representative.  But they tell that there is still a lot of potential left in this market.

Monday, April 25, 2005 (A.G.):
Autonomic Management: IBM Plays Bridesmaid Again

The same day (last week) that I was talking about the tremendous, mouth-watering possibilities of autonomic computing vis-à-vis data center management (using the Opalis interview as by backdrop) IBM, was pushing out some enabling technology for facilitating exactly such capabilities.  This was good and at the same time bad.

It is good that IBM, the undisputed pioneer on this front, is continuing to push the envelope.  It is bad that IBM’s efforts, in the main, are still focused on delivering enabling ‘toolkits’ (and in this case even a simulator) for other vendors, as opposed to delivering production-level solutions that customers can exploit today.

Yes, I know it takes time and money, but this is why we rely on IBM.  They have the resources and the motivation.  It is IBM’s large data center customers that would clamor for more automated management solutions.

The two new autonomic technologies [i.e. downloadable software toolkits], now readily available at IBM’s alphaWorks, are:

   1.  Policy Management for Autonomic Computing (PMAC), and

   2.  Touchpoint Simulator [i.e. a simulation test bed]

PMAC is very much in line with what I talked about last week and what Opalis is striving to deliver in terms of policy-based, unattended operation.  PMAC delivers "autonomic manager" functionality which when embedded within software applications will enable them to make intelligent decisions, on their own, based on previously specified policies.  Whereas Opalis talks about an automated database recovery/restart scenario, IBM (less ambitiously) describes a policy-based database backup scheme triggered by preset policies pertaining to the time of day, activity level or even vacation schedules.  I think we can still all work out the immense potential here.

I might be a bit harsh here (and some of you would inquire as to what is new in that) but the Touchpoint simulator comes across as being rather lame.  From what I can see it is a test bed for trying out autonomic concepts.  Yes, I will admit that this might be useful to some … but I am anxious to see real products not simulations.  But enough said.

These technologies dovetail into the Java-centric autonomic ‘architecture’ and toolkits that were unveiled by IBM in February 2004.  I did a detailed analysis of that launch along with some useful pictures.  At that juncture IBM also laid out a 5 level (or plateau) program showing how autonomic computing would evolve over time – with more and more of today’s manual tasks becoming progressively automated.  Level 5 (as shown here) was the last step of that evolutionary process.  Note that policy-based management is what defines Level 5.  With PMAC IBM is trying to get us there – but, to be honest, we still have long way to go.

You can find some introductory material on autonomic computing in the document section of Guruge.com.