cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefano Mazzocchi <>
Subject Re: 1060 NetKernel: virtual Internet operating system, XML runtime
Date Sat, 01 Nov 2003 10:11:39 GMT

On Saturday, Nov 1, 2003, at 07:53 Europe/Rome, Niclas Hedhman wrote:

> On Friday 31 October 2003 20:09, Stefano Mazzocchi wrote:
>> On Thursday, Oct 30, 2003, at 17:05 Europe/Rome, Rogier Peters wrote:
>> there is one say in english that I really like: "don't bite more than
>> you can chew".
> Instead of being negative, be inspired.
> 1. The Dependency based cache, where invalidation goes from leaf 
> towards root,
> could be useful to look closer at.

Dependency based cache is the way to go and it's exactly what we have 
in the 'eventcache' block.

There is a lot of work to do still in order to get dependency 
information transparently, but having the cache invalidation controlled 
by the resource-changing events, rather than from resource-requesting 
events, it *definately* the way to go.

> 2. The public/private separation currently in progress in blocks, is 
> probably
> a good thing.

sorry, I'm missing your point here.

> 3. Very little actually separates NetKernel from Cocoon. The fact that 
> more
> and more products are released with a Cocoon-like processing 
> mechanism, just
> adds to the total user base, and the Cocoon community does not have to 
> do all
> the "conversion".

Very true. The more people 'get' the pipeline concept for content 
creation, the easier will be for them to adopt cocoon in their 

> 4. I can imagine a Cocoon-compatible implementation on top of 
> NetKernel. Is
> that so bad? I don't think so, what about the day when W3C ratify the 
> Sitemap
> DTD as standard?

[sound of stefano puking]

> (I know you, Stefano, you will say it will be the death of
> Cocoon as we know it ;o)  )

My comments wasn't negative at the technological concept, but at the 
marketing around it.

I tried hard to make cocoon, despite its functionality, look "humble" 
from the outside. I have heard criticism about this, but I still 
believe that the reason why we are growing, slowly but steadily, in 
more and diverse markets (and in big companies!) is because we aim 
high, but we don't seem pretencious. We work with other technology, we 
don't aim at replacing them.

An XML operating system for the web is *exactly* the opposite of this.

People compare cocoon to... well... nothing. There are very few 
xml-based publishing systems. Now we are moving toward "web glue". How 
many "web glue frameworks" do you see up there? it's hard to 
compare.... might be hard to get in, but once you are in, those acronym 
maniacs strategical consultants will have a hard time understanding 
where cocoon fits... but since everybody seems to be using it or 
evaluating it, they will have to talk about it anyway.

Microsoft rule #1 for a successful market share: move the expectations. 
Microsoft is lame at technological innovation (they buy it from outside 
or clone it), but they are absolutely incredible at marketing 
strategies. [.NET, from that point, is already a success, despite the 
fact that nobody uses it in real life]

Marketing your product as "operating system for the web" will just 
throw you into a comparison with "operating systems"... the hardest 
market on the planet.

I know it's just a metaphore (they don't want to substitute linux with 
NetKernel), but it's a very bad one, IMHO.... and not even for techies: 
because they would expect much more from a "netkernel", than they 
actually get.

Also, the acronym soup around web services always gives me a very bad 

But don't worry, I never complain without taking a close look and see 
if there is something that I can take home with me ;-)


View raw message