[SWLUG] Not seeing the Code for the IDE

peter apvx95 at dsl.pipex.com
Sun Jan 30 21:05:58 UTC 2005


Dick Porter wrote:

>
>I don't know what you mean by "n-Tier", so I can't comment on that.
>However, I write C# code in emacs, use makefiles, and debug with gdb.
>(I'm part of the mono team at Ximian.)
>  
>
Hi Dick.  Thanks for the reply.

By n-Tier we mean that the application is broken down into tiers or 
layers.  For example, a web-based app will have:
*  A presentation layer (the html page)
*  A layer providing some logic behind the html (like the executable 
code in an aspx page or a jsp page)
*  A business logic layer
*  A data access layer

The first two layers will most likely run on the web server.  The 
business logic layer is encapsulated into remote objects that run on a 
separate application server (usually - because you can then provide for 
some security such as firewalling between the two servers).  The data 
access layer is (again, usually) located with the database on another 
server.

In principal, of course, all these layers could be on the same machine.  
Equally, they could be located halfway round the globe from each other.  
It shouldn't matter, and you should be able to locate them whereever you 
want when you come to deply them.

These applications have just too many files in them (usually at least 
100+) for me to be able to keep them organised without help.  And I hate 
building makefiles of my own: it's the sort of thing that computers were 
built to do.

You clearly don't have these problems.  I admire you for that.  'Cos I 
sure can't do it.  I want to be writing code, not worrying about how to 
compile it, or organise it on the file system.  :)

>As far as I can tell, the two useful parts of an IDE are the editor, and
>the documentation.  I've never yet found an inbuilt editor that comes
>close to vi or emacs, and the microsoft documentation usually sucks
>anyway :-(
>
>  
>
You are obviously better at remembering method names and overloads than 
I am: and you are also obviously a better typist.  I could never trust 
myself to get every call right first time (especially in terms of 
capitalisation, never remembering all the overloads and their parameter 
types).  Code completion means I never have to look for typos in method 
calls - which is a big saving for me.

You're right about microsoft documentation, though.  It's awful.  There 
are 3 things I miss in .NET that I loved in Java: one is the clarity of 
the Java documentation, the second is javadoc which is streets ahead of 
anything in .NET, and the third is Java exception handling.  Making 
exception handling a part of a method's signature is, in my opinion, a 
brave and entirely constructive thing.  You can compile your project and 
let the compiler tell you what errors you've missed.  Great for lazy 
programmers like me.

I really admire you guys who can get by without a sophisticated IDE.  I 
can't, and I'm sure - well I know - that I'm not on my own.  When I've 
got my dissertation out of the way, and if they'll have me, I'd love to 
try to help get monodevelop up to speed.  I'm convinced it's critical 
for mono's success.

Again, thanks for the reply.

Cheers


Peter



More information about the Swlug mailing list