[
Lists Home |
Date Index |
Thread Index
]
And that works because they are scripting to a
common object model made common by a common
browser. IE rules. That makes it work for
the non-nerd. Then when they want to go
a little deeper, there is no shortage of
examples starting with View Source. Cut and
paste and move on. Some will want Web Service
programming to work like that and with the toolkits,
it sort of does once past the legalities. But:
1. The average hacker isn't programming against
a db server. If he is and doesn't understand ODBC,
no amount of HTML is helpful. Sure, there can be
other kinds of DBs, but something is handling the
connection string, building SQL statements, training
the non-nerd about relational table schemas, child
tables, etc.
2. A web service past the hype implies someone
has worked out the terms of use in advance. It
may be a public site with laissez faire use, but
some kind of policy has to be in place. I think
these issues are the same without regard to REST
or RPC.
So I am back to wondering why REST over RPC. My
intuition is that average non-nerds don't build
web services without toolkits. If it comes down
to the toolkits, again, code is on the loading
docks as Dave says. I have to assume (maybe wrongly)
that all those global specifications (WS-Inspection,
Signature, etc.) are there for a reason and will
work reasonably well.
len
-----Original Message-----
From: Mike Champion [mailto:mc@xegesis.org]
2/18/2002 11:38:41 AM, "Bullard, Claude L (Len)" <clbullar@ingr.com> wrote:
>Do you really think the average office worker knows what the
><!DOCTYPE at the top of an HTML file is for? I don't. OTOH,
>they don't code them.
OK, I got carried away with the "average office worker" and
HTML ... but I do think that lots of non-nerds learn enough
HTML to tweak their homepages or weblogs, and simply ignore
(and don't touch) the stuff they don't understand.
I see my 9 year old doing it to tweak postings on a kids
website (neopets.com), FWIW, without any prompting or
explanation from me.
Likewise with URIs; I can more easily imagine an "average"
person accessing a Web service by tweaking a big ugly URL
and putting the result in a frame or a table or formatted with
CSS than I can imagine them buying VS.NET, installing some
libraries to make a SOAP call from Javascript, sorting
out when they must use XML-RPC and when they must use SOAP, etc.
Anyway, I offer this as only "the only scenario under which RPC
could decisively lose", not a prediction that it will happen.
|
