Re: SSD TBW

On 8/23/2026 11:44 PM, Lawrence D’Oliveiro wrote:
> On Sun, 23 Aug 2026 21:58:32 -0700, Don Y wrote:
> 
>> On 8/23/2026 8:47 PM, Lawrence D’Oliveiro wrote:
>>>
>>> On Sun, 23 Aug 2026 08:47:27 -0700, Don Y wrote:
>>>
>>>> And, its going to magically be able to compare JPGs, executables,
>>>> .so's, etc?
>>>>
>>>> Play with something like Beyond Compare (Windows) and see how
>>>> helpful file comparison can be.
>>>
>>> Does Beyond Compare do comparisons on arbitrary binary files?
>>>
>>> On Linux, for example, we can do something like
>>>
>>>      diff -u <(xxd file1) <(xxd file2) | less -iX
>>>
>>> to compare hex dumps.
>>>
>> Yes. If, for example, you have inserted a byte in one, it will
>> display a "hole" in that location in the other to show the balance
>> in agreement, though offset, in much the same way you would compare
>> text that had an inserted line.
>>
>> It tries to understand the forms of various file types and tailor
>> its comparison to that type. It's aware of many different formats
>> (excel, word, pdf, etc.) and tries to add value to each presentation
>> -- instead of just saying "they are different" (or "binary same")
>>
>> Additionally, the sources and destinations can be on remote
>> machines, accessed via SMB shares, FTP, etc. So, I can compare the
>> contents of one FTP server to another on a different host and note
>> the differences as well as synchronize them.
>>
>> And, it isn't plagued with the path length problem that Explorer
>> has, so you can use *it* to drill down deeper than Explorer would
>> otherwise allow. This is handy when one of the targets is a UNIX
>> machine and the other windows.
> 
> So really, what you’ve got there is one giant bloated executable that
> has to subsume the functions of what, in the Linux/*nix world, would
> be a dozen separate tools -- including the format-guessing functions
> of libmagic plus its own custom file explorer, it would seem. Only in
> a less flexible, more monolithic form.

It is extensible.  Ditto for CVS's handling of diff and merge.

What you (don't yet!) have in your UNIX approach is many executables
scattered around the file system, possibly not installed, and something
that tries to invoke each based on some criteria.

"Oh, gee, the pdfcompare program hasn't been installed on this machine.
Let me download a TRUSTED binary from a repo and try this operation
again."

"Hmmm, can't compare xwds on this box.  Maybe I can compare them as binary
files and try to guess the differences."

"Crap!  This version of XLScompare is buggy!"

How is that any better?