Having trouble finding a word in multiple files

Liam Proven lproven at gmail.com
Thu Jun 18 12:25:09 UTC 2020


On Thu, 18 Jun 2020 at 14:00, Chris Green <cl at isbd.net> wrote:

> Well, yes, compressed text is smaller than uncompressed text, but you
> have to uncompress it to search it (even if only momentarily).

OK. True. But the core point here is that _people do not know_ that
all the Office formats for the last 13 years are compressed. It's
internal and hidden.

And AFAIK this applies to Open/Libre Office, as well.

I have recovered a lot of lost/damaged/corrupted files over the years
with a simple search of binary on-disk data then some selective use of
``dd`` and ``strings``. This is no longer possible.

(People also don't know that LibreOffice's reverse-engineered MS
format important is more robust than MS' homegrown one. The easiest
way to try to rescue a corrupted MS Office document is to make a copy
and then attempt to open it in LibreOffice. It works fine more often
than it doesn't.

> Much of the advantage of plain text and/or simple markup language is
> that you can find things *with context* in them using grep.

Yes, absolutely, but only if you know how to use that tool, and only
if your OS contains such a tool as standard. Which Windows does not
and never has. Nor did DOS or Classic MacOS or Acorn RISC OS or Psion
EPOC or Novell Netware or any one of the other OSes that I am or was
once very familiar with and did hundreds of hours of work on.

One of the things Apple has grasped that most of the Linux world
hasn't -- although Ubuntu came close, before it killed off Unity, the
mobile products and refocused on server -- was that there is a huge
disconnect between the GUI world and the command-line world.

Most GUI OS users are totally incapable of using a CLI at all, and
those few that know it exists don't want it.

All the Unix stuff -- everything's a file, it's all just text, link
small single-task programs together with pipes to get work done,
script such actions for automation -- *none* of that applies to GUIs.
The GUI world is a whole different world of big monolithic apps which
use binary formats and only reluctantly share info by import and
export.

Some OSes accept that.

Classic MacOS had no shell, no concept of it, no config files, no file
extensions, no paths, nothing.

Atari ST GEM had no shell, but that's because it was removed -- there
were still config files, drive letters, file extensions, etc.

AmigaOS had both, but most users rarely touched the shell.

Windows (and OS/2 as well) was a GUI layered on top of a text-only OS.
NT isn't -- it's inherently graphical -- but it inherits all the
baggage from the old DOS versions.

The sad thing about Mac OS X for me is that it went backwards, to a
GUI on top of a text-mode OS, but it's a very elegant rich GUI and
most users never ever even open the terminal once.

This is still hard to achieve on Linux. Classic Ubuntu did pretty much
get there --  in the late GNOME 2 and Unity eras, with Synaptic, you
could get everything done and never open a terminal even once. But
it's moved away from that with the deep nastiness of GNOME 3 (IMHO --
I know others would disagree).


-- 
Liam Proven – Profile: https://about.me/liamproven
Email: lproven at cix.co.uk – gMail/gTalk/gHangouts: lproven at gmail.com
Twitter/Facebook/LinkedIn/Flickr: lproven – Skype: liamproven
UK: +44 7939-087884 – ČR (+ WhatsApp/Telegram/Signal): +420 702 829 053




More information about the ubuntu-users mailing list