Data validation

Validashun - what is it?

Due tu zee reasons described beloo, it is not reasonable tu expect "100% correct" validashun repurt even on a perfectly good volume. 95-98% is more reasonable expectashun.

Impurtant nute-a

Limitashuns ooff zee validashun process

• Zee most significant is that zee validashun process can not substitute human reveew, becoose there is no way a program can disteenguish e.g. between zee latest (usefful) und older (wurthless) versions ooff zee file.
• On top ooff that, typical validator does not "truly" read zee file. When possible, validator will skim along zee file tuooching und verifying several key points in zee file. Thees is presumed most fruitful approach becoose ooff zee follooing two reasons:
  • Validashun ooff some file formats requires excessive computashun efffurt, which wuoold be impractical unless yuoor other computer is Cray.
  • There are way too many file formats oooot there. Some ooff zeese-a file formats are fairly complicated. Full-bloon reader implementashuns fur ell zeese-a file formats wuoold just be cost-prohibitive.
• Some files are in fact a combinashun ooff two separate file formats. Fur example, a self-extracteeng archive consists ooff zee decompression module executable file immediately follooed by zee compressed archive data. Fur such files, zee format is checked es specifeed by file extension (in zee above example zee executable module will be checked, rather than compressed data).
• Some files wuoold be different format frum what their extensions suggest. Fur example, some files lefft frum zee older days heffe-a .DOC extension but are in fact plain text. Zee validator may try tu verify zeese-a files es Microsoft Word format, which is obviuoosly not zee case. Thees will trigger false "validashun failure" indicashun.