Back to home page

DOS ain't dead

Forum index page

Log in | Register

Back to the forum
Board view  Mix view

chkdsk recovery? (Users)

posted by Doug(R) E-mail, 11.12.2017, 20:15

And much depends on what type of files they originally were, how large each file originally was, and how fragmented the diskette originally was.

On 1.44mb diskettes, files are usually allocated space in multiples of 512 bytes (one 512-byte sector per cluster; other disk formats may be different). Files equal to or smaller than this will have been contained in one cluster, and will therefore be fully recoverable. Original files larger than this will have been contained in multiple clusters, perhaps discontinuous, and they might or might not have been left fragmented by CHKDSK's recovery process.

Text/ASCII files will be easiest to reconstruct, as they are human readable -- just import the segments into a text editor... and save onto a different disk! There may be "junk" at the end of the final cluster of any type file "fixed" by CHKDSK's recovery -- if text/ASCII, just delete it in the editor; if binary, it *might* be ignored.

If you don't want to try to identify a binary file manually (using a hexadecimal viewer), a great, up-to-date, file-identification tool is TriD, written and maintained by Marco Pontello:

Unlike other file-id utilities, the identification patterns used by TriD are contained in a separate data file, which is extensible and is updated on an ongoing basis (often weekly) -- the latest (2017-Dec-11) as of this posting contains definitions for an amazing 9,428 different file types. It has an optional command-line switch to automatically correct filename extensions, which you might want to use for the FILExxxx.CHK files.

The only catch is that TriD is a Win32 command-line executable that will require the HX runtime to run under pure DOS.

TriD won't be totally perfect in it's assessment of some .CHK files. Since text/ASCII files don't usually have standardized headers, it won't be able to accurately identify these (but you can easily do that manually). Also, .COM executables will most easily be identified if they start with a jump instruction (since they don't contain headers either). Finally, in the case of fragmented binary files, TRID will generally only be able to accurately identify the first chunk (where the header exists).

Hope that helps. Good luck....

- Doug B.


Complete thread:

Back to the forum
Board view  Mix view
15234 Postings in 1372 Threads, 251 registered users, 17 users online (0 registered, 17 guests)
DOS ain't dead | Admin contact
RSS Feed
powered by my little forum