By JGH, May-2006.
It is common for programs to put commonly used data in DATA statements which are then read into a set a variables at startup. A classic example is the names of months:

        DIM mon$(12)
        RESTORE
        FOR mon%=1 TO 12:READ mon$(mon%):NEXT mon%
        DATA Jan,Feb,Mar,Apr,May,Jun,Jul,Aug,Sep,Oct,Nov,Dec

This has a few immediate disadvantages:

• The data is held in memory twice, in the program itself, and in mon$()
• The DATA pointer is modified, unless LOCALised

You can avoid the DATA pointer being modified by doing the following:

        DIM mon$(12)
        mon$() = "","Jan","Feb","Mar","Apr","May","Jun","Jul","Aug","Sep","Oct","Nov","Dec"

but the data is still held in memory twice.

An alternative for small bits of fixed data like this is to hold them in a string:

        DEF FNmon(mon%)=MID$("JanFebMarAprMayJunJulAugSepOctNovDec",mon%*3-2,3)

This has several advantages:

• The data string only occurs in the memory once
• The DATA pointer is not affected

This can even be done for data that at first sight doesn't look like fixed data:

        DEF FNmonth(mon%) \
        \ =MID$("JanuaryFebruaryMarchAprilMayJuneJulyAugustSeptemberOctoberNovemberDecember", \
        \ VALMID$("010816212629333743525967",mon%*2-1,2), \
        \ VALMID$("785534469788",mon%,1))

The first VALMID$ string is a series of initial start positions of the month name strings for each month. The second VALMID$ string is the length of each month name.

Note: the example functions only give valid results for valid month numbers.