On Monday, January 28, 2002, at 08:11 AM, Yoel Jacobsen wrote:

> Andrew,
>      Thank you for your reply.
>      However, this only proves my point - your application fits 
> perfectly into 'playing with ideas'. Have you ever tried to open 2GB 
> database in squeak and make automatic analysis of the data? Another 
> application I needed - construction of LDIF files for customer's 
> existing data sources. It worked, but it was very slow and the image 
> was not fun to touch in the (long) time it ran.

Depends what you mean.  I have manipulated GB-sized files in Squeak and 
performed detailed analyses on it without cursing the darkness, yes.  I 
have no idea what you mean by "make automatic analysis of the data," 
apart from automating an analyis.  Frankly, I don't do much work with 
multi-GB datasets that aren't already in an RDBS.  The database, thus, 
takes care of itself.

>    I don't say I don't like Squeak. I liked it a lot. I just feel it 
> could be more suited to my type of 'real world problems'. On the other 
> way you can claim that my feeling is based on the fact that I know 
> Python better and longer. This might be true. What I wanted is to know 
> on what extent the readers of this group do I/O and memory intensive 
> tasks with Squeak.

It is possible that we simply have differing definitions of "real world 
problems."  I also knew Python better and longer than I knew Squeak, but 
once I moved over, I never looked back.  This said, notwithstanding the 
fact that I represent Zope.

>    I don't think this discussion is a troll, though.   Yes, I can 
> create a minimal image and run 'squeak minimal.image myscript' but it 
> would not give me the advantage of managing a running process in my 
> development image.

I have no idea what this last sentence has to do with practical use of 
Squeak.