Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/3/84; site sesame.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxn!ihnp4!qantel!dual!lll-crg!seismo!harvard!talcott!sesame!slerner
From: slerner@sesame.UUCP (Simcha-Yitzchak Lerner)
Newsgroups: net.micro.pc
Subject: Re: Re: Re: Norton Utilities (Unerase, actually)
Message-ID: <210@sesame.UUCP>
Date: Tue, 23-Jul-85 00:15:06 EDT
Article-I.D.: sesame.210
Posted: Tue Jul 23 00:15:06 1985
Date-Received: Sat, 27-Jul-85 04:10:38 EDT
References: <185@drivax.UUCP> <18923@mgweed.UUCP> <440@nbires.UUCP>
Organization: Lotus Development Corp
Lines: 40

> Now I am really getting confused!  I thought that the directory information
> on the disk contained the location of the first allocation block for the
> file described by that directory entry.  Once you know the first allocation
> block, you can use the FAT as a linked list to find subsequent allocation
> blocks.  I assumed (probably incorrectly) that when a file is deleted, each
> FAT entry in this linked list is set to zero (000) to indicate that this
> cluster is unused and available.  If this linked list is destroyed in this
> way, then the file would be unrecoverable (I think).  On the other hand, if
> this assumption is incorrect, then DOS must at some point do "garbage
> collection" to recover space used by deleted files.
> 
> Can anyone out there in NETland (Microsoft, are you listening?) clear up my
> confusion about this? Thanks.
>


To clarify, your assumptions are correct.  The directory entry does 
point to the first entry in the fat.  when a file is erased, the first
letter of the file name is overwritten with E0h and the entire linked
list in the fat for the file is zeroed.  When dos needs new space, it 
scans the fat sequentialy for zeroed out entries.  There is no list
or pointer to available space.  (I think to many cp/m hackers have
been putting in their $.02 and confusing things.)  by the way, a
fat entry points to a cluster, not a sector.  a cluster is a group of
sectors, # of sectors per cluster dependent on volume type. (SS floppy
has 1 to 1 correspondence)

Hope this has clarified the issue.  **sigh**

 
-- 
Opinions expressed are public domain, and do not belong to Lotus
Development Corp.
----------------------------------------------------------------

Simcha-Yitzchak Lerner

              {genrad|ihnp4|ima}!wjh12!talcott!sesame!slerner
                      {cbosgd|harvard}!talcott!sesame!slerner
                                slerner%sesame@harvard.ARPA