Re: signing off

[Alan Shutko <ats@acm.org>]

Chris Majewski <majewski@cs.ubc.ca> writes:

> For me, the "hard" way usually turns out to be the easy way in the
> long run.

Not always.  There _are_ problems with Linux (and other free Unices
and even commercial Unices, to some degree) and it's important to
recognize them and address them.  You just can't demand that one
distribution fix all of them immediately, and when judging
distributions you should guage it against others, not always against
Windows.

For example, printing is truly a mess under Linux.  If you have a
postscript printer, have all apps generate postscript, and don't use
any special fonts, you'll find it fairly easy to get going.  However,

* Any developer wanting an app to print has to learn Postscript,
  because there are no general APIs to use.  (Note: This is being
  addressed, albeit in a fragmented way.  Different libraries are
  being put together to make it easier to print.  Yay!)

* Any developer wanting to print with user-installed fonts or printer
  fonts needs to write their own way to access those fonts, because X
  doesn't give you outlines or all font metrics, and you can't get any
  info from Ghostscript.  This means users have to install the same
  fonts in tons of places.  (This is being addressed, but in a very
  fragmented way.  It's going to get way worse before it gets better,
  since different groups are trying to solve it on their own, or are
  just shoving it away and hoping it goes away.  Look at the miserable
  font/print handling in KWord, Star Office, WP (two COMPLETELY
  incompatible methods depending on version), and I won't even get
  into TeX.  Additionally, defoma is trying to be a way to
  automatically install said font into all those different places, but
  it's still a mess.)

* Retrieving printer information to optimize output (resolution, paper
  size, etc) is nigh impossible.  Some apps (but not all) can now
  parse PPDs, but most of the printers installed don't currently have
  PPDs (because they aren't postscript).  

This isn't even getting into what happens when the print job hits the
spooler, which is somewhat less of a problem but is so unstandardized
that it vacuums SAN from anyone trying to support the different
spoolers.

Sorry... this is one of my hot buttons... I just want to point out
that indeed, there are things left to address.

-- 
Alan Shutko <ats@acm.org> - In a variety of flavors!
You never hesitate to tackle the most difficult problems.