from Charlie Reuben Dallas(Home of the Meyerson Symphony Hall,the US' BEST) RE: All XBASE Problems in the PC and the Cost to fix.(25-50 Billion???) Part I Intoduction to the PC X BASE Problem: No one on these Year 2000 study groups is looking at the real problem for the PCs. As they say every 4 years here, "Its the APPLICATIONS,dummies!!! Not the programs. It was what was done to those programs that have built up a mess that is a scaled down version of the MAIN FRAME Business Application Problems. And...its worse because there is some agreement on what you can do to isolate and identify and fix the Big Boys. There has been little said or done that even indicates that ANYONE is even looking at the most wide spread mess of all, the X BASE problem. The question of the PC/MAC/Small User Problems vis a vis The Year 2000 Problem has not been quantitated properly because the problems have not been evaluated in depth. Given the magnitude of the Global 2000 woes, this is understandable. There has been much press on the Gartner Study claiming 300-600 Billion for the larger Enterprise fixes. Little has been said of the micro-bombs in the Micro Computers (the PCs/Mac etc). Using conservative extrapolations from several data sets, I arrive at a minimum cost to Small Business of a range of 25 to 50 BILLION Dollars (with a mean of 37.5) to get these houses in order. I "believe" it will be double that range if properly measured maybe more. Much will be passed off as "upgrading" because for small businesses there are tax advantages that large Corporations can not utilize if they "upgrade" just for Year 2000. (See Yo' Countant). I define Small Business as businesses with less than 100 employees. The US Census says there are 38 Milllion of these. The Dallas Chamber of Commerce says there are 76,000 in Dallas County Alone(2 million souls ((mostly all perfect Angels)) ). If ten per cent of those companies spend only $10,000 "fixin" (Texas Tech Speak) that be...76 million. Lets look at the 38 million. Same thin,Lucy.... $10,000 X 3.8 Million. Let me help you...3.8 Mill X 10 =38 mill X 10 = 380 Mill X 10 = 3.8 BILLION X 10= 38 BILLION. Look at it from the MOST elemental way. Dallas County (Chamber again) 1-4 employees: 41,947 firms. Is it reasonable to assume that a 'business person" might be willing to spend $10,000 on Computers? If she knew she had to file electronically and the computers made her money??? Forget all the people who "cant afford it" and think only of the 10% who WILL PAY.. I think it quite reasonable that a small 1-4 person company would lay out $10,000 to get working in 2000 but I will hedge my bet knowing that 1/2 of those will be history by 2000 and so I'm really betting that 1/5 will pay that kind of money. (Two desktop machines a laptop, software and some 'var" help). If they already have the machines and software and choose to "modify" the apps. who gwine do it??? The VAR who wrote the Air Conditioning Service Software and only charges $5,000 a copy?? Or FTD who charges every Florist X dollars to use their Software etc. Its a true "pay me now or pay me later " deal. So considering ONLY 10% of the very smallest Companies in Dallas County we have: 4,195 X the $10K we still have almost 42 million. Plus TAX. (Dallas County pays for its Year 2000 problem from that alone,I hope).I will calculate this another way later. At the heart of the cost will be the modification of or migration from any Customized Software Applications. Not the BIOS problem and NOT technically defective Software. Recently, some efforts to catalog the Generic Software have been made. They MISS the Central ISSUE. For, it is in the modification of the Generic Software (whether it be a Language or a "type" of software such as XBASE or SpreadSheet) that small and mid Range and even Enterprise level that Year 2000 problem willl arise. I do not think that the cost of the X Base problem has been included in most of the Year 2000 studies, yet the applications infest many large companies and must be fixed. Given the nature of Corporate Non Disclosure policies in most Companies, such proprietary information is not "given out". I can only think that if there are 8 million "seats" for Lotus "NOTES" there must be 5 times as many customized "things" sitting in the Big Companies. In small companies, they don't even have a clue yet. NONE. Zippo. In X BASE...there is NO Global FIX..NONE.. I repeat: there is no GLOBAL FIX and...there are zillions of problems. Zillions is a technical term of a Matrix of millions of application programs multiplied by a matrix of millions of users (Note: its really a TENSOR but lets not get that technical). The Ultimate Night Mare of the I.T./I.S. Pros came true....'The USERS GOT LOOSE". Unlike the "Big Guys" the Little players don''t know yet. AND.. worse, their problem will be more costly and possibly not fixable. WHY?? Simple. The Big Guys Problems are really analogous to that of a Boeing or Lockheed Jet craft user problem. The Airlines have internal experts, external consultants, Manufacturer's consultants and expertise. More Brains and Bodies can be hired. At the other end of the spectrum, there are the car owners of the PCs with many Makes and Models all "customized" for the users, with many different kinds of driving styles and paint jobs. The Computer is the 'Universal Machine'. Give it a problem and a way to do something and it will work on it. It is "silly putty" for brains. Soon it will "replicate" and truly make us humble as it does winning at Chess. But...in the PC world you have all these machines and all these programs but NO FACTORY TRAINED MECHANICS Certified for Year 2000. There are NO University Courses in Year 2000 and prior experience is not on the job discription. It never happened before. What's a toy maker or a chain of dry cleaners to do??? There are only "hints" in the Heloise like columns of the PC mags. NO real 'sink your teeth into the code" solutions. To know the X BASE problem, you have to be able to read the code and that is not your Mother's bedtime story. AND..you really have to know whether it "can" be fixed. And..you have to know whether or not you can get the "data out" in a meaningful way. It goes without saying that you must back up everything. All financial stuff is needed for the you know who anyway. IN its original form. (maybe or maybe not,see your Accountant). In the very worst case scene comes the compiled software for a nice little company whose VAR retired. The Source code is gone to the Dumpster in error. The applications all have 2 YY date displays and ALL manipulations in the machine see only 2 YY. These are the Cinderella Class : "DEAD DUCKS". Gonzo. They will fail or lock and the data will not be retrievable. Think it can't happen. Ask any computer person what "losing the pointers means". If the program looses its "reference points, It either stops and locks up or loops forever or simply "gives up" (crashes). You boot it up again with your back up and same thin... Lucy. but worse because now the backup is gone. Can't happen??? Take it from me the proud owner of $20,000 worth of zeros and ones when a "proprietary accounting package" lost its pointers. On a less stressful note, the BIOS problem of the Micro Machines themselves is well known. A new retrofit chip or machine may solve that problem. IMHO,a better safer fix for many will be the Software Patch written by Tom Becker (formerly of Dallas) while he was here in Dallas. Becker's program is available at www.RightTime.com or by link from my home page sites. Chip upgrades across large applications should be done by professionals. Because no machines were really "Y2K" enabled much before 1996 (save for the claim of Gateway) and there is an installed base of 100 Milllion or so machines the choice for businesses is simple. Install a new Bios, use Becker's solution or enter the date in January of 2000 and pray. Lets say that only 1/3 the installed base choose to use the former two. At a cost of $50 per fix plus labor at $50 (you clearly can't let Steve Salesman do it because he spills coffee) thats only some 3.3 Billlion. Where do the rest of my mean 37.5 Billion get spent? PART II Consider the X Base Problem. There are 8 million in the dBASE installed base per Borland's numbers. dBASE's Data base "engine". Lets triple that to arrive at a modest amount of copies of data base software in the field including "compiled applications and double that again to reflect the multiple number of applications per desk. MOST if not all of the 'custom" software applications use date field information in some manner just as COBOL applications do. Since dBASE,Foxbase, SuperBase,Paradox and Clipper amoung others were "easy to work with" readily 'adaptable" and had the data base at their core, these were often the center of development for VARS and "business partners" (B.P.s).aka: "Solution Providers". There was also a great deal of Basic at first and then C and now of course, C++. (JAVA is next). In the "off the rack" application department there were programs like ACT and Quicken that were built one way or the other that were not Year 2000 Compliant (both are now according to reliable reports and migration is easy). Technical flaws in other well known programs have been corrected to implement Leap Year and Year 2000 4 YYYY changes. For all practical purposes, these were probably not mission critical applications (except that a lot of business could be lost if data in a Contact Manager were lost to a sales or Marketing person.). Again, migration and labor might cost $100 (though if I had a site license for something that was defective I personally would want it for free hint hint to Gates and MS). So doubling the 3.3 Billion again is probably in order. Where's the other 30 Billion??? X BASE... and all those wonderful '"integrated" customized programs that are literally everywhere.Some even wrap all the business applications of one industry for a company. and the "VAR" was clever. He made his "solution" scaleable so the company with 1,000 employees is using the same mess that the ones with 50 or 100 use. The big bird paid for all the others and the rest was sheer gravy. Many,many of these were written in dBASE III,IV OR FOXBASE. In many cases, the programs were implemented in the mid or late 80s and are still working their hearts out. ALL in mm/dd/yy. The biggest adventure for these was to move them to Windows 3.1. (such a thrill I had to pass on until 1995). And..note,,,it is NOT Borland's fault. They built a perfectly wonderful machine (for me anyway) and its been "compliant" since when..1985??? What's the problem?? Consider the following, in ALL the XBase engines there is a choice of settings for the display of the time and date. The US standard (to our regret now) has been MM/DD/YY. Has anyone you know EVER entered the year 1996 in a program except for a letter to your Mother in a Word Processor??? USERS are fools. :Programs need to be FOOL proof. Whence the 2 YYs. Why give the fools a chance to enter 1896 or 1886. Besides, in the Mid 80s those 2 extra YYs added up and since they were all 19 why bother. As long as the "Engine KNEW".what was the harm. Enter the problem stage left. IF you enter any date to 12/31/99 the machine sees Dec.31,1999. Enter 01/01/00 and the MACHINE SEES Jan. 1, 1900. THAT is critical. Proof of this is simple and proof of what the machine sees is simple. Enter 02/29/00 and you WILL GET...."invalid date" or some such message..BECAUSE there was no leap year in 1900. Test further and enter leap year 1904 and it goes down smooth. Now go back to the command line function in your data base (if you can) and write {SET CENT ON} and proceed as above. The leap year 2000 will go in and all seems well. So...to the LAYMAN or even the non PC Computer Guys the next question is "What's your problem? Do a "global fix"." AND that is EXACTLY what people on the Net Mailing lists think is the end of the X BASE problem. It is not. Part III: A Case Example: Consider: Super Widgets, Inc. makes 1,500 line items of stock widgets and customizes some 5,000 more for special customers with 500 Manufacturer's reps sending in orders by fax,phone and telex. Sometimes one of the more advanced uses Email. The process begins. Somewhere along the way, Sam Super decided to "computerize" so at the Country Club bar he found out from the bartender that I.M.A Schlocker specialized in Small Computers for "important business men like Mr.Super". I.M writes up a full scale integrated package for all of Sam's departments and pretty much does a bang up job using Zippy BASE. (I love Borland so I don't want to say dBASE or the Foxy word). Many of them look a bit like this: ------------------------------------------------------- ---- ** foolish.prg ...prints a list of foolish things business men have to do ona given date; ** written in Zippy Base by I.M.A.Schlocker 4/1/87; ** Property of I.Schlocker and Co. Call for special programs at my day job at 7-11 (212-555-1212); ** Modified by E.B.Schlocker II 4/1/92 ; "------------------------------------------------------ ------------" clea all; set colo to y+/b; set century off; use foo inde foo,foo2,foo3; do evnmofoo ; @ 20,20 say "Printing a list of today's items for you"; repo form foolist.frm ; clos all end;(if evnmofoo was a routine) end; return NOW...the part I want to stress is that for each and every little one of these here guys, inVARiably there IS a date field and inVARiably , the neighborhood VAR set the Cent off (just in case) to make sure that some two fingered typist didn't have the chance to muck things up much. Never did anyone exit a program and set Cent ON. So, you have in every possible way shape and form, applications written "to order" for some "entity" or "market" each with its own peculiar way of doing things, names of fields, methods and procedures which internally depend upon those names and data types to produce whatever was supposed to come out the other end of the machine. appplication TWO: mostly P Code: order widget issue ship orders issue invoice post A/R monitor A/R call for check when late post A/R paid post G/L if we just look at: issue invoice: ship stuff shipping notice and or bill of lading terms: 2% ten net 30 days 1.5 if over 60 days retro to 30. (pretty standard) integrated with A/R Here we see the need for the use of 'memory variables mem vars" If those Mem Vars depend on a 2 YY date field and can not read the 4 YYYY field we need to rewrite this application. And..we now open up an entire can of worms,especially the more tightly knit the entire "integrated package" was built. Since every name and field may have some use in other programs and modules, tweaking Super Widgets nose can make his feet hurt or even drop off completely. This is what your Pascal teacher was trying to tell you about "goto"s. Now that we have "objects" and OOP maybe this kind of thing will be gone forever. I don't wager on such things. Nobody every lost money betting on the bad judgement of people. If we instead choose to "migrate" we need to know that NOW. because for MOST Companies...there is time to "convert/expand the date fields" so it can be "pushed into" the new application,tested and fielded. Soon, that window will not be there and for the PC users there is NO TIME. NO HELP and NO precedents for this kind of thing.. Who you gonna call???"Data Busters"???? Part IV: The Costs Who has to "do this" fixin or migration to new software ??? Well, just like the Big Birds. YOU do. There are two ways to calculate this. ASSUMPTIONS: 1. The cost of buying new software, modifying it and migrating data and users to it will probably be the same as "fixin" (Texas tech-speak) the OLD or MORE. 2. The cost of inventorying, analyzing, "fixin" and testing/fielding old increases as the number of users and size of the company increases. This is probably exponential or sigmoidal. Small companies with less than 20 employees and a "modicum" of expertise would do well to consider migration and burying the cost as a "capital improvement" (see your Accountant). This is not an option for large companies who must expense Y 2 K "fixin". ROUGH GUESSTIMATES: APPROACH ONE: Assume all "iron" stays in place. Cost of Bios,mandatory replacement of obsolete software (old versions of ACT or Quicken or 123), plus modification of code of vertical Xbase Software. $1,500-2000/per machine/user. Stir in 'user downtime" administration costs" the whole bit. I think that's "Cheap" if it represents 1/2 hard costs and 1/2 soft. What's it cost to "train" a user??? I'm cheap because you just give me the stuff and I use it. Joe Stumble in sales needs what??? $500 in training. CompUSA did $350 Million in 'Corporate Training". They charge to teach you how to use what they sell you. If "the boss" teaches, in a very small company what's that time worth???? In actual billing or sales time vs. lost time? Lets use the widely published "# of PCs used in business". : 100,000,000. world wide. IF...only 1/4 do anything about Year 2000 plug in the numbers. and LO and Behold...some sort of range: 37.5 to 50 Million. (Note: the per centage of people using PCs for business who will do something is greatly elevated from the 10% given at the start of this paper because for any PCs in ANY serious sort of application in a Big or Mid Range Company there is "no question" that they HAVE TO DO IT or risk down time in production. Period. That would imply that 90% or more of any PCs in Corp.hands would be made Year 2000 compliant. So the only difference here is the ownership of the machines.This further implies that the cost estimates given are probably far too low.) APPROACH TWO: New Iron and Software. New programs. Try... the SUN MICRO numbers. A PC costs 10-12 K per year for a big Company. What is the "incremental" cost if you have to Consider Year 2000 because you are "forced to"?? You have the same costs as above in "soft costs" Plus you have the extra cost of "migrating what has to be migrated" and storing what has to be saved for the IRS and Your Industry Standards. Lawyers may save "forever". Doctors and medical too. So, is the "incremental cost" forced by Year 2000 reasonably the same as above or more??? Given the trick of "depreciating it" by making it an "investment" instead of household monthly cost it probably works out the same but see Yo' Accountant). (SIDEBAR..I'm real good for a Real Estate person. I send people to Lawyers and Accountants two groups known as "deal killers" to the Unwashed members of my BusinessTribe.) So again, we have the same results more or less (probably more)...a mere 37.5 to 50 BILLION DOLLARS. All to be spent unwillingly for something that we can't even moan about or vote on (like Congress). Best, Charlie Reuben, Charles P. Reuben BS,MA Hancock Properties DALLAS 8600 NW Plaza@Hillcrest S:3-C Dallas,Tx. 75225-4210 214-369-9502, or f:369-9337 E-ME: texasrltr@aol.com Year 2000 Links http://members.aol.com/texasrltr/buytexas/index.htm (temp.pages) Member: GDAR,TAR,NAR, and ..the Gt.Dallas Ch.of Commerce