Year 2000 Issues

Is your research data at risk from Y2K?

Aron Roberts

If you maintain research data on a microcomputer or workstation in which dates are an important component, your data's integrity could potentially be at risk as a result of Year 2000 (Y2K)-related computer problems. In some cases, the software you use in processing and analyzing your data might also be at risk of returning inaccurate results due to such problems.

Various sections of the campus's Year 2000 website at y2k.berkeley.edu can help you identify and resolve such problems. If you find that this Web site fails to fully answer one or more questions you may have about how to find and resolve Y2K problems on your computer(s), or if you find anything on the site confusing or difficult to follow, please let us know as soon as possible. Send your requests and comments to the feedback email addresses listed on each of the site's pages.

Based on the recommendations offered on the Y2K website to date, the following overview lists questions you might ask to ensure that your research data, and the processing and analysis of this data, will not be disrupted by Y2K-related computer problems.

Key areas of concern

Is your data being backed up? If you're not already doing so, now is the time to begin regularly backing up your computer's vital data, including research-related files. It's best to always retain at least two or three prior generations (i.e., snapshots taken at a particular time) of any valuable files. You might also consider making two data backups during the latter part of 1999 and setting these aside in a permanent archive. One of these backups should be made just before you start to find and fix Y2K problems on your computer, in case anything should go awry during that process. The other archival backup, which you might consider making late in 1999, would allow you to revert to that backup should unforeseen problems occur during the Year 2000 or beyond.

For some suggestions about how to back up your data, see Ask Dr. Micro: How can I back up the files on my computer in the April-May 1997 BC&C, and Simplify your life with the UCBackup service in the Winter 1998 BC&C.

Is your computer integrated with research equipment? By following the "Research Equipment" link from the campus's Year 2000 website, you can find both general recommendations and specific protocols and procedures that can help you evaluate your research equipment for potential Y2K problems.

If you have a microcomputer or workstation that is closely integrated with research equipment, check with your equipment vendor to determine whether any Y2K-related fixes to that computer are required. At least a few vendors may be willing to replace or upgrade computers and software which were supplied as integral components of their equipment, while others might offer some degree of support for computers used with their equipment.

Otherwise, if your computer has a data acquisition add-in board, specialized scientific software, or other such products installed which are integral to your research, you will need to determine whether these are Y2K-compliant. For guidance, see the "Other potentially important areas of concern" section of this article, below.

Are you using customized applications? If you or others are using application programs that were "written from scratch" or extensively customized, these may be at significant risk of harboring Y2K problems.

If an application used to acquire, process, or analyze research data was written in a scripting or programming language integral to a data acquisition product or statistical package; developed using a widely-used programming language such as Basic (including Visual Basic), Fortran, C, or Perl; created in a database environment such as dBase, Clipper, FoxPro, Paradox, or Microsoft Access; or built using client/server or rapid application development tools; it's important that you evaluate the application for potential date-related problems.

By following the "Customized Applications" link from the campus's Year 2000 website, you can find a detailed, phase-by-phase process that can help you (1) identify what work needs to be done to make your customized application Y2K compliant; (2) determine how best to get this work done; and (3) oversee the process of fixing, testing, and implementing the application.

Other potentially important areas of concern

Does your PC's internal clock/calendar have a Y2K problem? The internal clock/calendars in many industry-standard PCs may revert to 1980, 1984, 1900, or some other year when the Year 2000 arrives. Although this problem can be easily identified and resolved, it could impair gathering of research data if left uncorrected.

For example, if your PC is recording dates and times when sampling data, or is running scheduled processes, these activities might be adversely affected if your computer's internal clock were to revert to an incorrect date.

Are your application programs Y2K-compliant? Certain off-the-shelf application programs, including some popular statistical, mathematical, database, and spreadsheet programs, are known to have Y2K problems whose scope and impacts may range from trivial to very serious. If you're using one or more of these programs to process or analyze research data, it is important that you determine which of your programs may have Y2K problems. You can typically do so by viewing product Y2K compliance statements provided on your software vendors' websites.

It might be possible to ignore or work around a few minor problems in certain application programs, but in nearly all cases you'll need to patch, update, upgrade, replace, or retire your programs that have Y2K-related glitches. Many vendors provide no-cost, downloadable Y2K fixes for their recently released products.

Are you using stat or math programs? A category of applications likely to be of particular concern to campus researchers are statistical and mathematical programs. If you're using programs such as SPSS, SAS, SYSTAT, Matlab, Maple, or Mathematica, either on a microcomputer or workstation or on a larger system, we suggest that you check the web page Campus Researchers: Specific Y2K Concerns (http://y2k.berkeley.edu:7040/computers/fixpcs/issues/researchers.html), which is accessible via the "Computers" link on the campus Year 2000 website.

This page provides links to UCLA's Office of Academic Computing, which offers detailed information on Y2K issues in SAS, SPSS, and Stata; to Indiana University's Stat/Math Center, which has summarized the Y2K compliance status of some 30 statistics and mathematics packages and subroutine libraries; and to Academic Computing Services, a group within IST's Central Computing Services department which can find and fix Y2K problems in SAS and SPSS applications, as well as in some C and Fortran applications, on an hourly fee basis.

Do your data files contain date-related problems? Some data files may also harbor date-related problems. Examples of files which could potentially be at risk are spreadsheets and database files containing dates or performing date calculations. Certain of these problems might first surface when the Year 2000 arrives or when dates falling in the Year 2000 and beyond are first used.

There are a variety of tools and techniques that can help you identify date-related problems in your files. Some areas to look at when evaluating data files for Y2K risks include:

• Custom macros, scripts, and user-defined functions. Some spreadsheet and database files may incorporate custom-written macros, scripts, or user-defined functions which improperly handle dates.
• Standard date functions. Many spreadsheet and database programs provide standard date-related functions which are known to have Y2K-related "usage issues," such as Microsoft Excel's =DATE(), Microsoft Access's =DATESERIAL(), and Lotus 1-2-3's @YEAR().
• Dates stored as text or numbers. Dates stored in spreadsheet cells or database fields as text (e.g., "02/21/15") or integers (e.g., "990405"), may present risks when these values are converted (via standard date-related functions or otherwise) to dates, or are compared or sorted.
• Data exchanged via text files. Data stored in or exchanged via text files often may contain dates with century-ambiguous two-digit years. An example is "05/13/29", which might reasonably represent either May 13, 1929 or May 13, 2029. Such dates could be variously interpreted as falling either into either century depending on the rules used by the different programs which may import data from these text files.