Netscape integration

[Steve Lamb]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Wed, 30 Jun 1999 23:22:13 -0400 (EDT), Ralph Cohen wrote:
>It's actually very simple, I use different editors for different jobs. 

    Yet you could use one editor, one set of keystrokes for muiltiple jobs
and be more efficient.

>Also, in my business, we
>use a lot of specialized scientific/medical terms and jargon, which I
>doubt would be included in any generalized dictionary, so I expect to
>have to add certain words to almost any dictionary I chose to use.

    Then use a specialized dictionary.  You can make one or other people can
as well.

>Not crap at all.  There are options included with PMMail speller which
>tailor it for use with email.  Further, I demand different things from
>that speller than the one I use with my word processing software.

    All of which could be used since the aspell would be a *LIBRARY* they use
in place of the current one they have now.

>Would it be easier or more efficient for Bob & Ike to fix someone
>else's' code than it would be to fix their own?  Would this be a good
>use of their limited time?

    You just don't get it, do you.

    Aspell is free.  It is more powerful & more accurate than the current
spell checker.  In fact, it is more powerful and more accurate than most
spell checkers out there now.

    It has taken the Aspell team a few hundred man hours to get it to that
level.  For Bob&Ike to get something as power, as fast, as flexible, it would
take them at least the same amount of time.

    But it would take them only a dozen or so to get PMMail to use it.

    Now, here's the trade offs.  As Aspell is updated, since it is LGPL'd,
those code updates are released back to the public.  This means as it
improves, PMMails spell checking, since it is linked to that, would also
improve.  What would the time Bob & Ike spent on improving that portion of
the code.

    0.  Nothing.  Nil.  Zip.  Nada.  None.  Nul.

    Also the Aspell team, and anyone else who is interested in the code, are
not only constantly working on making it better, they are also squishing
bugs.  
Time B&I spend doing that.

    0.  Nothing.  Nil.  Zip.  Nada.  None.  Nul.

    *IF* there is a bug that comes shows itself in PMMail and B&I spend, say,
an hour fixing it, they have to submit that fix back to the Aspell team for
inclusion in the next version of Aspell.

    So, they spend an hour here, and hour there fixing *possible* bugs in the
code.  Let's say, oh, 15 hours to port and 5 hours of Aspell specific bug
work.  That's 20 hours (and that is being liberal, I think) of work.

    What do they gain?  First they gained the couple hundred hours of work
because they didn't have to try to recreate the code.  Then they also get
dozens of hours each release since they don't have to update the code for
improvements.  Then, on top of that, as new dictionaries are available, they
can ship those as well.  American English, British English, German, French,
Spanish, Medical.  I've seen all of those for Ispell and they will eventually
make it to Aspell.  There's a few more hundred hours of work.

    So, let's call it a nice round 300 hours for them to recreate the work
pluss another, say, 30-40 hours each year in upgrade, not counting new
dictionaries.  The now have a net gain of 310-320 hours with more each year.

    310-320 hours could be enough to at least write a decent IMAP
implimentation.  Maybe not great, or earth shattering, but usable.  

    Do you understand now?


>Sorry, I have enough trouble getting timely responses from Bob/Ike for
>PMMail problems without expecting them to be effective advocates on my
>behalf to other developers.

    That is a problem with Bob/Ike then.  You can also contact the Aspell
developers directly and ask them to fix something.  If they do, it gets back
into the next upgrade on the PMMail side and Bob/Ike spent time working on
something else.

>Nope,  I'm not panicking, but I don't see what the Linux efforts have
>to do with PMMail.

    Because you're are so indoctorated into the "Who do I sue" mentality that
you don't see the benefits in just man hours.  By legally using LGPL'd code
they can improve their product thrice over with a minimal effort and will get
incremental upgrades at no cost to them.  By adopting the Aspell libraries
they have the Aspell team working for them.  I'd like them to use the PCRE
libraries as well so PMMail can finally get regexps, something no other
Windows mailer does and would make the filtering system stand out again.  

    All they need to do, legally, is return any bug fixes back to the
community and make the library code available for download.  Personally, if I
were them, I'd do it, end up with a better product, and donate a small
percentage of the proceeds of PMMail back to the authors directly or to other
OS ventures.

>Once again, after seeing the quiet abandonment of the PMMail list by
>SouthSoft, I don't have a tremendous amount of faith in their ability
>to field an effective infrastructure for tracking bugs in PMMail as
>well as third party components.

    This, again, is a problem with Southsoft, not with the idea.  

    Of course, which PMMail list are you refering to?  Southsoft never ran
"this" list.  It was originally started, as it was this time, by someone who
loved PMMail enough to start a mailing list about it.  It was never offically
Southsoft's.

>Very impressive, but unlike you, my responsibilities do not include
>maintaining a public server and further, this discussion concerns
>applications and not operating systems.  Nevertheless, when I looked
>around for a site on which to host my company's web site, I chose one
>running Linux and Apache for exactly the reasons you stated above.<g>

    But it does concern applications as well.  The point is the speed of
development cannot be matched in a commercial venture.  If there is a bug in
the Aspell code, the Aspell team, someone else with the code, or Bob & Ike
could nail it.  The point is, right now, PMMail code has two people looking
for bugs in it.  Bob & Ike.  The Aspell code has a dozen or so people looking
for bugs in it.  

>>    You call bug ridden, bloated, barely functional, useless feature laden
>>programs a "success"?  

>Your characterization, not mine.

    Then what would you call, say, Microsoft Office.  Netscape.  IE. 
Windows?

>happens to work reliably and efficiently for me.  I certainly don't
>fault you for your efforts to shape PMMail into the type of email
>software you'd like to be using, but I think you'll agree that your
>requirements may not be the same as most other email users.

    No, they aren't.  My requirements far exceed most other email users. 
However, many of the propsitions I'm making aren't an "either/or" matter.  It
isn't "either they work for me or they work for you, but not work for us."

    Take Aspell.  It is a library that Bob and Ike can used.  If they use it,
it frees up time for them to concentrate elsewhere.  They can keep the
current interface and features (by deciding what to run through the
libraries, just as they do now) except PMMail's spell checker is vastly
improved and will continue to be improved.

    Now, how is it going to a more powerful, robust spell checker somehow
deminishes PMMail for you?  It improves it for me, you'd most likely not
notice a difference.

    How about moving PGP external again and putting in support for GPG.  It
would prevent the GPG crashing bug I have now, something Bob & Ike cannot do
since they don't have the code to the SDK.

    If they do that, it frees up their time to concentrate on other portions
of the code to put in both features I want and features you want.

    Same with the editor.  Most people here, I'll wager (in fact I know given
the ignorance show in the discussion) have never used the external editor
dialog.  People are so hung up on the internal editor except, when you come
down to it, the behaviors are almost identical and given maybe 10 minutes
most people would adjust quite nicely, esp. if the editor came with PMMail
and was preconfigured in the default installation.

    So, here's the question.  If PMMail is nice now, how much nicer could it
be if a good 1/2 (at least) of the work that Bob/Ike put into it no longer
needed to happen and you suffered, at most, *ONE* minor interface change?

    Imagine double the effort into core of the program, how it behaves, how
it handles the databases, what type of servers it can connect to, etc.  

    It can happen with an increase in performance, an increase in
reliability, and increase in power that average joe users won't notice the
difference on and power users will drool over and thank their lucky stars.

- -- 
         Steve C. Lamb         | I'm your priest, I'm your shrink, I'm your
         ICQ: 5107343          | main connection to the switchboard of souls.
- -------------------------------+---------------------------------------------

-----BEGIN PGP SIGNATURE-----
Version: PGPsdk version 1.0 (C) 1997 Pretty Good Privacy, Inc

iQA/AwUBN3rrBXpf7K2LbpnFEQJzXgCgzcZd8fREjUvB2Cn2uoEZ8QNaRlAAoKD7
WkKi1GHv5S6t9Cb9cYTjC8BJ
=lAJe
-----END PGP SIGNATURE-----