On Tue, 05 Jul 2005 04:11:29 -0700, Avi Bryant <avi.bryant at gmail.com>  
wrote:

>> In just about any professional desktop development tool today, the  
>> process of displaying DB data in form or grid format is nigh  
>> automatic.  I can do it without a line of code, and I can do it even if  
>> I've never seen the tool in my life.
>
> And then what?  I'm not challenging your statement,

No worries. I keep using Squeak because I love Smalltalk and Squeak has a  
lot of good uses and a lot more potential. There are certainly support  
communities wher the response would be "love it or leave it".<s>

> I'm just curious: say you use one of these tools to display a grid of DB  
> data without a line of code.  What's the process to then have a finished  
> application?  Do you have to write some code at some point?  How far  
> towards your final goal does that code-less grid take you, and how far  
> do you have to then go with code?

In some cases, pretty damn far. Some people don't want much more than an  
easy, slap-down, look at their data.

Personally, I've always had a problem with "data-aware" controls, in that  
they muddle the presentation with the data layer and often make the actual  
programming you have to do--the "business rules", I think is how it's  
currently phrased--heavily bound to both. (A lot is being done to improve  
this, though progress seems slow to me.)

However, it's easy to underestimate how powerful a demonstration it makes.  
I'm involved in a project which is based on an Access-like tool that  
really, really shouldn't have been. But the ease of creating and  
displaying data sold the people I work with. And why wouldn't it? Out of  
150 screens, 120 are forms or grids based on tables or sets.

It's been disastrous precisely because the tool--which does this one thing  
quickly--is stuck in '80s era DOS-style programming.

> In my personal experience, the tools that work best are those that  
> either let you go all the way to where you're trying to get to without  
> *ever* writing code (obviously this only works within a restricted  
> domain, but it's great when it does),

I've never gotten very far with that. IBM's VisualAge Smalltalk could do  
that, and was pretty cool, though.

> or those that acknowledge that you're going to have to write code at  
> some point and so focus on making that as easy and productive as  
> possible.

No disagreement.

> Whether or not I have to write "a line of code" to achieve the first 5%  
> of my goal tends to be vastly overshadowed by whether I have to write  
> 1000 or 10000 or 100000 lines to accomplish the other 95%.  I may have  
> more to show in VB than Squeak after an hour, but if it's a month long  
> project I know which one I'd want to use.

Yeah? You think that the absnce of a solid grid component might not eat up  
a bunch of that month? If not...well, what are your plans for the rest of  
July?<s>

> It does make for nice demos, though.

Which should not be underestimated.

> At any rate, you're certainly right that Squeak is not a "professional  
> desktop development tool".  If you're looking for one of those (and, I  
> gather, on the Windows platform), try Dolphin.

Well, we were discussing "general acceptance" and, in my case, of what  
people come and ask of me. It's not a matter of me "looking for a  
professional desktop development tool". (Personally, I doubt I'd use any  
Windows-specific Smalltalk.) I will be looking for a replacement  
environment/tool soon for the abovementioned big app, and I suspect we'll  
end up using Java. (I'm not excited at the prospect, but it could be much  
worse.)

And, you know, it's =fine= if Squeak is never meant to be a "professional  
desktop development tool". If Squeak has a fundamental "failing" it might  
be that it tends to excite the desire to make it all things to all people.  
If I can use it to teach kids and experiment with cool things, that's cool.

It's just not "general acceptance".<s>