[Seaside] Re: Why - XHTML perfectly good names are nowgiven en Seaside new names

[Norbert Hartl]

Fritz,

On 18.02.2011, at 01:31, Fritz Schenk wrote:

> Boris, I appreciate your comments. Yes, it is possible to 'improve' via the open 
> source model.
> 
> Let's play the entering anchor game.
> 
> Start with 
> 
> html anchor
> 
> Nothing is offered here with the code completion - we have no idea what is next. 
> We look into WAAnchorTag and discover that an instance variable of 'url' exists. 
> An then... 
> 
> You might browse that callback: is indeed available as a possibility.
> 
> Is a better code completion the tool we 'old timers' need?
> 
> What got me interested in Seaside is continuations as I implemented a 
> continuation capable controller in VB6 more than a decade ago.
> 
> I put up with the way of Seaside and suffer through its 'handling' of html 
> button onClick: ((html jQuery expression: 'external_links a')
> onClick: 'return confirm(\"You are going to visit: \"+ this.href)');
> 		 with: 'Attach Click' 
> when I notice Seaside's complete lack of knowledge about JavaScript when it 
> thinks that onClick: 'return 
> confirm("You are going to visit: "+ this.href)');
> 		 with: 'Attach Click' 
> is complete and correct?
> 
> Oh, well - I still like Seaside mainly because of Smalltalk.

you are right regarding code completion. I think we all can agree that it would be a good thing to have a code completion tool that is capable of what you want. But we don't have it at this time. I'm sure it will be there not far from now but... And if you talk about the early nineties than you probably couldn't imagine back then that there could be such thing like autocompletion. And with your knowledge about HTML you couldn't achieve much nowadays. And you wouldn't know what javascript is at all ;)
The html canvas is an _integration_ of two technologies: HTML and smalltalk. So there are some alternatives how to implement them. You can make one technology look like the other (making seaside methods exactly the same like HTML) or you can try to enrich both sides with their counterpart. Seaside is the latter one and you prefer the first one. Well, that's it, it is just not like you expected it. And there is no reason to change it if you listen to what others are saying. Furthermore you can take Boris' advice. Overwriting canvas and add your convenience tags is a no effort.
Bringing javascript into the mix makes them three technologies to intermix. Here the integration is much more complicated because the integration target is far apart. So you can inject pieces of javascript code into the stream of html and see what happens. That is a common way of doing it. If you use SQL then the statement is not checked before you send it to the database. It is evaluated remotely. And that is the same here. Even In javascript itself (in the browser) you don't have the ability to check the code without doing an eval (which is the compiler). Beside the syntax errors there is a big bunch of other problems where you would need to take all your injected javascript parts etc. 
I don't want to argue that this cannot be improved. I'm just wondering from where you take your expectations.

Norbert