What's the point? That the .Net framework is as big as a big RB app?
Don't understand, since you need to download the framework only once.
And this is something that we've needed in the RB world - a way to
deliver a generic RB framework that can be installed and made
available to all RB apps. While the total of an RB app would still
be as large as it is, what the customer sees in a download will be
much smaller on an application instance. I took a look at one of my
Mach-O UB RB tools and found that if the framework was globally
placed and published on the system, the resulting download from the
user perspective would be 7.5MB instead of 13MB (MBS and rbframework).
So, while the RB frameworks only add between 5.7MB and 10MB to each
app (unlike .NET's 20MB+), we wouldn't be populating users' systems
with potentially hundreds of copies of rbframework.dylib (at the least).
This is never the good idea it seems to be. Global libraries and frameworks
introduce fragile dependencies in applications and pretty much guarantee
that applications will at some point either not run, or not run correctly.
Granted, all of our applications rest on OS APIs and frameworks. They pretty
much have to in order to interoperate with the host computer and have the
look and feel end users expect. But anyone who has been programming for any
amount of time will immediately recognize that this to is a fragile
dependency. Changes in Mac OS and Windows APIs have broken countless
applications.
A global RB framework would change, what, 4x a year? There's no OS support
for keeping this up to date like there is with Cocoa and .NET. A binary
requesting a specific version would likely not be able to run or when it's
needed most. A less discriminating binary, one that could use different
versions, would be more bug and crash prone. An improperly written installer
could wipe out the foundation for every other RB app on an end user's drive.
DLL hell for RB.
If an end user is using numerous RB applications, that person is still going
to end up with copies of the framework, a dozen or more in 3 years, if their
apps are updated. If a person uses one RB app that gets updated regularly
that person will also end up with multiple copies even though they have one
app.
For what? To try and save space? Under what scenario do you actually save?
And how much? How many people realistically have 100 RB apps? I've got my
entire music library, thousands of digital photos, and numerous movies/TV
shows, all on one drive, and I still have 100 GB to go. If I had 100 copies
of the RB framework at 10 MB a piece I would have...99 GB to go. This is the
drive that came with the computer. As an end user I would much, much, MUCH
rather have a program just work, without a lengthy install, then worry about
a measly 10 MB. Or 1 GB. Disk space means nothing to me. RAM means
something, and so does execution time, but disk space is meaningless now
days.
Daniel L. Taylor
Taylor Design
Computer Consulting & Software Development
[EMAIL PROTECTED]
www.taylor-design.com
_______________________________________________
Unsubscribe or switch delivery mode:
<http://www.realsoftware.com/support/listmanager/>
Search the archives:
<http://support.realsoftware.com/listarchives/lists.html>
