[REBOL] Re: Some thoughts around Rebol
From: rgaither:triad:rr at: 13-Apr-2001 14:25
>Hi there, while reading through most of the postings and observing RT
>progress to a complete product strategie I'm always thinking about some
>stuff, which is needed to make Rebol a "professional" tool (that a big
>company is going to use it):
This is going to be a hard thing for REBOL.
1. Big companies buy "applications" not technology platforms.
2. But, they buy those "applications" only for technology platforms
they have heard of - unless no other solution exists. :-(
3. There has been some serious consolidation in the "application"
product space that leaves big companies only buying from other
big companies. It is getting harder, and harder for a vertical or niche
application to make it big in those companies.
4. Even in the application "Hosting" space it is a hard sell without
dropping the name of the technology under it - Oracle, Java, ...
>1. Storage Formats
>I would like to see RT to adapt XML as the generic storage format for all
>kind of data. Rod Gaither just answered to Carl why he is going to use XML.
>I fully agree with Rod. To use Rebol in a corporate environment it is
>mandatory to support "standards" and using XML seems to be a good way.
Thanks for the prop. :-)
The XML support should be visible, but not a replacement. It is too big
to do all the way and too much of a distraction to the REBOL core goals
to make a central aspect. It does need to be WAY better than it is with
the current product though which barely counts as an after thought. :-)
>2. Object serialization
>I really like the functionality to store Rebol objects to disk and read them
>back in again. With /Command we have access to normal database systems. What
>I think would be a killer-functionality for Rebol is the following: Being
>able to store objects as XML (yes, we can do it but native support would be
>nice) within an object-oriented XML aware database, which supports the
> - handles huge number of objects (several hundred thousands)
> - fast searching (quering)
> - handling of object-types as properties
> - keeping Rebols run-time memory footprint low
>Imagine that you can handle a lot of objects in a way that Rebol will only
>load the stuff as needed. At the moment we have in memory-databases, that's
>OK for some hundred objects but not for a enterprise application.
This is such a tough issue. I too have wished for everything I have in my
rdbms but with XML flexibility. Schema defined on the fly, with binary storage
and retrieval speed, built in support for all the ways of traversing and selecting
based on XLink/XPointer/XPath - all in a small, simple to use for big or tiny
projects... :-) Way too many smileys to be a reality discussion!
Again, this is a product waiting to happen, but it is not REBOL.
>3. Office Access Layer
I won't touch this one.
I'm trying very hard to escape the clutches of this beast!
I'm down to using less than 10% of any of the office product
features. It is time for simple, functional, open replacement
>That's it for now. Feel free to add your thoughts, flame mine ;-)) ...
No flames here, good thoughts in general.
Oak Ridge, NC - USA