Phillip Pearson - web + electronics notes

It seems that everyone using CSS eventually resorts to using absolute positioning with widths specified in pixels, to get everything to work. How, exactly, is this better than using tables?

Clarification: I hate using CSS for layout. It seems unrealistically hard to do simple things in CSS. Good results are possible, but take far too long to get right. CSS for fonts, styles, borders etc, on the other hand, is great. Convenient, straightforward, and easy to get working.

A friend just popped up on MSN and asked me how to generate S19 checksums in Python. It's a pretty trivial bit of code, but I know I'll want it again, or maybe someone is out there googling for a way to do it, so here we go:

def s19_checksum(line):
    length = int(line[2:4], 16)
    bytes = [int(line[i*2:i*2+2], 16) for i in range(1, length+1)]
    return ~sum(bytes) & 0xff

for s in ('S1130170707172737475767778797A7B7C7D7E7F03',
 'S111FE7B59597B0173F6014A7B0146B7573093',
 'S106FE891B823D98',
 'S109F8020107DF01BA0D6C',):
    print "* checksumming %s" % s
    print "  checksum: %s (compare with %s)" % (hex(s19_checksum(s)), s[-2:])

Update: Incorporated Michael Urman's suggestion. Previously I was using eval('0x' + line[i*2:i*2+2]) to parse hex numbers. Thanks Michael - I didn't know that int() could do that!

Update 2: Fredrik Lundh points out a) that Python has a sum() function to add up a list of numbers (better than my reduce((lambda a,b: a+b), bytes)), and b) that s19_checksum() calculates 0x4D for the last string, while the checksum included is 0x6C... this is actually the reason my friend contacted me. He had received a hex file with a suspected bad checksum (the 0x6C) and wanted to check it out :-)