[linux-elitists] What's some marketing buzzwords for what we do?

Rick Moen rick@linuxmafia.com
Thu Jan 2 14:13:28 PST 2003

Quoting Larry M. Augustin (lma@lmaugustin.com):

> OK, I'll try the new SO6.
> But I'm wary of the "new release is a completely different beast" argument.
> It sounds an awful lot like something MS might say.  (As in Windows 95/98
> sucked, but WinXP is a completely different beast.)

I can say on the basis of my experience with OpenOffice.org 1.01 that
SO6 is likely to strike you as incrementally better but still sucky in a
few fundamentals.  (You're one of the few people on this mailing list
who's likely to seriously exercise a word processor before returning to
preferred text-processing tools, so I'll be interested in your report.)

It has all the advantages and disadvantages of a one-binary-does-everything
application.  (The illusion of separate modules is just that, and I'm
told that all of its various document formats are really all the same
XML-based document format.)  It grabs 73 MB of RAM for its initial load.
By comparison:  AbiWord 1.02, 6 MB; KWord 1.1.1, 17 MB.  Many operations 
are a bit sluggish.  The HTML export format remains an atrocity, if like 
me you still hew to the notion of avoidable dependencies on
physical-attribute markup being bad.

Where it really shines is in reading Microsoft Corporation's binary
formats.  I've heard that recent iterations (at least) of MS-Word,
MS-Excel, etc. write out as their document format what amounts to a
snapshot of their memory state.  Which explains why they're so difficult
to parse, even by Microsoft Corporation's own code.  In fact, I've heard
of a number of cases where Microsoft's applications ceased being able to 
read a word processor document or spreadsheet that they'd created, and, 
before discarding the thing as corrupted, someone had tried StarOffice /
OpenOffice.org, successfully parsed it in, written it back out again,
and suddenly Microsoft's apps _were_ able to read the document again.

In other words, SO/OO.o is really good at parsing MS's formats, and is
much more conservative than Microsoft itself is, in how it writes such
documents.  (I've read comments from the programming team, about this,
and they warn that because of their code's conservatism in
document-creation, SO/OO.o may cause documents' size to increase a

Also on account of the one-binary-that-does-everything design, SO/OO.o
is uniquely qualified to deal with Microsoft compound documents.  If
someone has embedded a spreadsheet fragment in an MS-Word document, the
odds are that SO/OO.o will deal with it, without fuss.

But that reminds me of an anecdote:  My mother-in-law mailed me one of
her recent grad-school papers, for my review.  She offered it in PDF; 
I said just send your MS-Word original.  AbiWord parsed it mostly OK,
but there was some binary gibberish that hadn't parsed cleanly,
including most of the bibliograpy.  This was peculiar (and
unprecedented), so I researched further.  OO.o had similar results, as
did WVware.  I asked her to lob over an RTF version, for comparison:
Pretty much the same results.

Now, most people would not have gone for one or another of two surmises, 
without doing that much legwork, depending on their inclinations:
Either the results meant that no tools on Linux could yet read MS-Word
files reliably, or the original document was subtly corrupted in a way
that only MS-Word could still deal with.  

But, as it turned out, upon asking a few key questions about the muffed
text in question, it wasn't a pure MS-Word document at all:  It seems
that MS-Word's handling of bibliographic information is sufficiently
inept (compared to, say, WordPerfect) that many academic users and some
business users install ISI Researchsoft's EndNote[1] as a third-party
"add-in" module to MS-Word.  The data that AbiWord, OO.o, and WVware
couldn't parse fully was a binary EndNote insertion into an otherwise
ordinary MS-Word document.

OO.o is, in short, an essential tool, if you're trying to be maximally
compatible with Microsoft office-productivity tools.  It's just not
instantly _likeable_ the way AbiWord and Gnumeric are.  And it sometimes
fails to be fully compatible because the document's creator has
absent-mindedly created a document using third-party tools that rendered
it non-standard, and neglected to mention the fact.

[1] http://www.endnote.com/

Cheers,           "I don't like country music, but I don't mean to denigrate
Rick Moen         those who do.  And, for the people who like country music,
rick@linuxmafia.com         denigrate means 'put down'."      -- Bob Newhart

More information about the linux-elitists mailing list