Mailing List Archive


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [tlug] [OT] Japan and 海外 credit cards



On 2009-08-03 23:46 +0900 (Mon), CF wrote:

> > Yeah, I hear some banks are still using OS/2...
> 
> or Excel 97 for that matter (it's 2009 now :-)

Actually, I'm not quite with you on that one.

Excel hits a particular sore spot. It may be starting to change, but
in recent history, give any quant (referring to a quantitative analyst
in the financial domain) a nice shiny new copy of Excel 2007, with its
cool multi-threaded recalc and all, and watch him recoil in horror.
They radically changed the user interface for what seems no apparent
reason except to use up yet more of your valuable screen real-estate
with menus. And they provided no way to get the old interface back.
Everybody I know went straight back to Excel 2003; who needs the pain?

> It's not as simple as "saving it in the latest Excel format". The  
> customized *.dll excel plugins to perform very complex calculation  
> formulas written in the MFC4.0 (Win95) days all needs to be ported  
> to .NET or whatever. You can do it but you need to be 100% accurate,  
> right to the exact dime and penny, at first try.

I disagree here, too; what's the point of being so accurate when your
old spreadsheet was wrong in the first place? I'm extremely impressed
at the sorts of things people do in Excel, but also live in fear of it,
and the stories of people spending huge amounts of time debugging their
spreadsheet ports only to find subtle problems in the old spreadsheets
are legion.

Looking at other programs, Word happens be something I was, back in the
Word for DOS 5.0 days, quite intimately familiar with, since I assisted
in the generation of about 15,000 (at least that was the part I saw)
pages on contracts for building the stations on the Jubilee (subway)
line extension in London back in the early '90s. It wasn't difficult
at all to work up a good set of style sheets and train the other word
processing operators in their use, and things went very smoothly.
Gone were the days of the WP ops using preview mode and inserting and
deleting blank lines throughout the last 150 pages of a document after
an edit in order to get rid of a widow or orphan line on the second-last
page.

Well, until we upgraded to 5.5, which brought in the new CUA interface
(with those now-familiar drop-down menus and dialogue boxes). Things
started to degrade from there. Word for Windows came in before I'd left,
the style sheets got even harder to use, and we quickly reached the
point where we are now, in which you basically say, "screw the style
sheets; just make it look somehow more or less like it should through
any means necessary, and hope we never have to edit it again. Also, make
a PDF from the output, because we know the formatting is going to go to
hell as soon as it gets on a system with slightly different fonts."

Resumes and short contracts tend to make a great example documents for
a WP system, since they are short enough to be typed in quickly, and
have formatting requirements that are not entirely trivial, but nothing
terribly hard, either, if you've got a good system for setting up style
information and tagging paragraphs with styles. Someone who knows the
concepts and knew a different word processor or two could sit down with
Word 5.0 for the first time, type in the content, set up a dozen styles,
tag the text for it, and have a decent looking document that's easily
reformatted within two or three hours. Word for Windows and OpenOffice
have defeated me on this every time.

It's so bad that I, someone with non-serious needs who formats perhaps
one document every few months, finally gave up and spent more time than
I'd cared to learning and mucking with LaTeX because it was less work
than battling with those damn "user-friendly" programs.

OpenOffice, by the way, has in my mind demonstrated how badly the open
source community can fail; they've basically replicated all of the
problems of the Microsoft Office suite and added some more of their
own, since their main goal seemed to be to keep the user interface that
creates most of those problems in the first place.

> And the cost and time involved might be unjustificable. So all in all
> you're still stuck with low productivity regardless of how great the
> new version of Excel is .

Well, the new versions are just not that great; that's the issue.
They're not working on fixing the problems, mainly related to the user
interface (this goes far beyond what's on which menu) that lead to the
errors.

> And then we have COBOL, the super hated super resistant tech  
> cockroach. :-)

Don't worry; I anticipate that by the time the Y2036 problem rolls
around, most of this will be ported to Java, which I like to "the new
COBOL."

Back during the rise of TCP/IP and its battle with ISO/OSI, Marshall
Rose wrote a book on ISO/OSI subtitled, "The Protocol of the '90s." That
subtitle applies more perfectly now than ever. I think Java will remain
"the language of the '90s" far into the future.

cjs
-- 
Curt Sampson       <cjs@example.com>        +81 90 7737 2974
           Functional programming in all senses of the word:
                   http://www.starling-software.com


Home | Main Index | Thread Index

Home Page Mailing List Linux and Japan TLUG Members Links