Two Possibilties -- Part II
By Mark:-
2. It's not manageable. (using traditional approaches)
.....Looking at the current state of Y2k efforts, the scope of the problem and the interdependence all businesses have on each other, it's easy enough to argue that this thing is spiraling out of control.
.....If this is true then Y2k is a fright train barreling down a one way track toward a passenger train coming the other way.
.....IBM is issuing disclaimers saying they are not liable, lawyers are briskly rubbing their hands together as their firms organize Y2k legal divisions, big six accounting firms are visibly back pedaling as they try and untangle themselves from Y2k completely.
.....Budgets are growing daily, we are not shocked to hear numbers for individual companies that are 8 and 10 digits long. Salaries for talented individuals are skyrocketing while entire IT departments are being gutted as head hunters scramble to find capable people and once found are doubling their salaries.
.....If these are signs of anything they are signs of a crisis out of control.
.....Possible solutions? I don't see many, possibly a cooperative effort and worldwide mobilization on the scale of ww II. But who will take charge? How will they, if they even can, organize everyone involved? Maybe the federal government will do it, but when? Probably it is better left to private enterprise.
......So lets assume for a minute that such a mobilization will happen. Virtually every business along with every Y2k company teams together and decides to cooperate, share resources and focus talent where it's needed. Basic economics says that someone needs to be paying for this project and of course everyone expending effort needs to make a profit. I'm a dreamer but not to the point where I think charity could rule on this, it's too big a
problem and too much is at risk for greed not to be the biggest motivation to insure cooperation. Another impetus is business survival, at this point the panic button doesn't seem to have been pushed. Is profit sharing enough of an impetus to motivate everyone.
......How does such a system structure itself? Do we see one very gutsy company standup and strong arm everyone else into agreeing and cooperating, or is everyone (Viasoft, IBM, etc...) scared enough by the visible lack of progress that they are already talking and following a basic human characteristic of crisis management. This may sound far fetched but is still in the realm of possibilities.
.....Does this Y2k partnership (not yet a reality that I can see) say, stop! Quit assessing and re-assessing the impact and get on with the important part, actually fixing the problem.
......This is where I see the factory process (yes MatriDigm) being a big player, if (granted we have yet to see much proof, but the people involved have made some bold statements) so, if MatriDigm has a highly automated tool that can guarantee 100% of the findable dates can be found and fixed at a high rate of speed then why would it not become the equivalent of a Normandy beachhead?
......We've heard the one million lines per hour claims, suppose they can really get close to that, at this point I sure hope they can, because so far very little code has actually been fixed and as front-end work proceeds the backlog of code ready for conversion should be growing exponentially.
......TSR(?) recently said they can convert 400,000 lines per week (I think it's per week) but even if they can do that much per day what's involved in getting the code to the point where it can be converted? Their press release seemed to indicate that a lot of pre-conversion work was done. Peritus is said to be able to do 300,000 lines per "team" per month (again this is from memory and subject to being off somewhat), how many teams are
possible? what conditions are required? But the point is we are looking at a problem so large that these numbers quickly become insignificant in the big scheme of things.
......If MatriDigm has what they have claimed to have, and if the factory can actually process code at rates nearing 10 million lines per workday, what's to stop them from building as many factories as are needed? If they perfected the factory conversion process to a point where it could process the majority of COBOL programs out there with minimal human intervention what would prevent them from building say 10 or 20 factories, allowing them
to convert upward of 100-200 million lines of code per day.
......I realize this "scenario" requires strong faith in the technology and people behind this company, and I realize that delays, and a lack of contracts has dampened many peoples enthusiasm, and provided the skeptics with an unending thread of criticisms. But I also realize that what they have developed needs time to be trained and tested to the extreme. I'm not worried at all (well almost) that we have seen no iron clad contracts so far.
With the slow start Y2k has gotten off to, and with the chaos that is found on the front-end (lost/missing source code, poor documentation and so on) it's no wonder little code has been converted to date. I believe it's true that MatriDigm has not missed its window of opportunity on this. And in fact, looking at the progress in Y2k so far I believe that MatriDigm is in a stronger position now then at any time in the past.
......What has to happen now, and what I believe will happen is a crisis intervention group will form, this group will consist of a partnering of all the existing Y2k companies and probably many fortune X companies. Resources will be pooled and directed to maximize effectiveness, the time consuming tasks such as triage, picking and choosing what's vital and what's not will be minimized and systems will be converted almost en-mass. The first
such conversions will happen in phases allowing extensive systems testing on the back-end, but as successful tests prove the technology phases will be eliminated and testing minimized to the point of being a minor part of the process. I can almost imagine entire massive applications being converted and put back into operation in a matter of weeks.
......Right now time is being spent on front-end, as well plans are being implemented for detailed system testing. System testing is needed on a grand scale only because traditional fixes introduce so many errors and it is unthinkable to put an application on-line without rigorous testing. Testing is needed on some scale, but as code run through the factory is proven testing will be minimized, it's a matter of proving the factory and its
ability to accurately do what it's designed to do. How many successful conversions will be needed to prove the technology is really the big question. At what point will CIO's and IT folk determine that the risk of error has been all but eliminated?
......Grand thinking? You bet, believe me I realize that all too well, but with just the superficial knowledge I have regarding Y2k much less then this will not suffice. I've researched and followed this topic for several years and I am sure that without something on this scale occurring soon many information systems cannot hope to be functional come the turn of the century.
......Lets say MatriDigm can convert all the COBOL that can be brought in, and also that MatriDigm can and will add additional language support (PL1, RPG, DB2 etc...) I don't know how big these markets are (I don't think anyone does) but estimates from Gartner Group are still standing, and in fact being raised. The liability issues are said to exceed one trillion dollars on top of that.
None of this may happen, but if things continue as they have for the last year regarding Y2k in general I will be very worried indeed. Traditional approaches to Y2k are failing miserably, Front-end work is dragging along at a snails pace and in fact being redone over and over as management tries to get it's arms around the scope of the problem. Something has to give soon.
Something will give soon.
Mark
Foolishly seeking knowledge
.......the revolutionary way. |
