Here's an exchange I had with a guy in the MS Site Server general Usenet Newsgroup... can anyone find exception with his or my arguments? Additionally, he doubts the ability of Zope to handle large sites (near the end of the message). Any clarifications are appreciated....
On the opposite, ZOPE is interpreted, not only the templates, even the server itself. No paradigm like COM is part of the game (though it might be extended). So ZOPE will be less speedy.
COM happens to work just fine with Zope :^) I've written a demo ExchangeFolder before that let you march into the Public Folders on your Exchange server as a little test too see if it would work. To me, the whole argument of "COM components are compiled, so they'll be faster" is sort of a lark anyway. If you wanted to, you could crank up your copy of VC++ and crank out a COM component for your business logic. The fact that people don't seem to be doing this (or even clamoring for info on how to do this) is telling. From the point of view of a web developer, the issue is not "gee, how many milliseconds faster can this COM component execute my query", its "how fast can we get this project done". Some people seem inexplicably not to have caught on to that yet :^)
Also: I really doubt the scalability of the ZOPE Object database, compared to the SQL Server that is part of the site server
licensing.
The ZOPE Object database - well - there are no benchmarks on it.
There are no benchmarks, true. It might be worth noting, though, that the actual backing storage of the object db is pluggable, documented and extensible. This architecture is specifically designed to allow the use of alternate backing stores (oracle, other enterprise dbs, etc.) -- including SQL Server if you were so inclined.
"When you do this, you're basically extending the Zope environment by creating your own object types without having to know a high-level programming language (ala C++/Visual Basic COM object development)." - well, just you do it in a slow interpredet language, where C++ COM Objects will be speedy!
Yep, speedy as hell - and a year late ;^) See above.
"The power of this model is formidable. For example, I developed a "news item" object in 22 minutes which allows my users to add "news items" to their department for perusal by other department members." - nive. What about adding 100.000 news items. What is the performance then? This is unrealistic? OK, lets make an interface for a news server. 1 million messages a day. Youre dead.
This seems a simpleminded argument. If you are going to do something like a news server interface, you are also going to do the obvious design work beforehand that takes scale risks into account. I believe he entirely missed the most important points of your statement - "22 minutes", the implied department-level scope of your solution, and the fact that it worked for you. The "youre dead" bit makes me wonder whether it's really worth much time bothering with this guy - you can't really answer someone who obviously already knows all the answers :^) Brian Lloyd brian@digicool.com Software Engineer 540.371.6909 Digital Creations http://www.digicool.com