Re: Longhorn Killer

[Helge Hess]

On 22.03.2004, at 15:38, Rogelio Serrano wrote:
> Foremost on my mind is a "winfs" done right. Call it "LinFS" if you 
> like. That means a storage system that does not use sql.

SQL is a great query language ...

> I have been trying out domino notes lately and i think it would be 
> cool to use the notes data model for a desktop storage system.

... which of course can also be used on a Notes database. If you don't 
like SQL syntax, use DASL syntax or whatever, but the basic idea is to 
select a set of object properties and to filter and sort on them as 
well as being able to query on arbitary subtrees of a hierarchy.
Which is pretty cool for managing certain kinds of data.

Indeed OGo includes something like that (project document database), it 
uses a combination of NSFileManager and EODataSource as the Objective-C 
API, so this would be possible for GNUstep. But of course using a 
database has quite some overhead in case you want to do just regular 
filesystem stuff.
Eg the gain of being able to sort the whole harddisk on filetype "gif" 
in milliseconds certainly isn't worth the loss of performance due to 
the increased management overhead on any regular operation.

> Of course search would be very much a part of the desktop right? I 
> have also been thinking hard about smtp and spam and security lately 
> so maybe a pgp distribution system with a ui like notes would also be 
> a good idea. make it easy for people to make pgp certificates and 
> distribute it.Im trying out the idea right now. I was writing code to 
> implement some kind of nsf and forms and views. I was going smoothly 
> until i read a paper about main memory databases and now im kind of 
> stuck trying to implement that. Another is cairo and X Windows done 
> right. I like Keith Packard's idea. I want to try xcb with gnustep 
> back. I dont know how much work is needed there though. I stopped 
> working on Y when i went to the freedesktop.org website and really 
> read almost evrything in there. I think this is true for all linux 
> desktops. Another is an installer system. Im sure you can add more.

Yes, all that, including the storage system, are all good ideas, but 
none is a unique advantage of Longhorn and is going to kill anything 
else. You shouldn't buy all the buzzwords how cool everything is going 
to be - Microsoft did not deliver something which actually increases 
productivity a single time in the past ;-)

In contrary, when Longhorn is due, it will again be an *excellent* time 
to migrate users to other systems, because people will be told to buy 
upgrades. This worked very well to the advantage of Linux in the past.

> Well I admit you make me made stop, look and think very hard.

Excellent ;-) Its always much better to think hard and explain before 
prior starting such threads on a list ...

> I asked that question because i thought somebody already has an 
> answer. I might have a wish list but im not an expert in GNUstep. I 
> just wanted to start testing as soon as possible.

My impression is that GNUstep is mostly about reimplementing OpenStep. 
It is especially not about going after buzzwords like .NET or Java. 
Longhorn is about buzzwords.

You seem to be a lot after buzzwords, so consider:

> I have not done any deep analysis of KDE+GNOME at all. I have used it 
> and its nice actually. Gnome and KDE have gone a long way since but i 
> have never really been able to go over the learning curve to actually 
> build programs that I want. I just dont like the complexity of these 
> environments.

Well, Longhorn is also a complex environment. It won't be much easier 
there, it just might have better tools.
I guess GNOME+KDE are much more in line with your ideas than GNUstep. 
Especially in the GNOME/Mono mailing lists you'll be able to find 
plenty of people sharing your "vision".

> I came to like GNUstep after i read the ProjectCenter tutorial. The 
> environment seems less complex to me.  Maybe because its incomplete 
> but i felt that a lot of nice applications could be written in a short 
> period of time with the current state of ProjectCenter and GORM.

Indeed, but this has a completely different focus. I guess someone who 
loves Longhorn is very unlikely to consider GNUstep an option and the 
reverse.

> Its all about going beyond unix.

IMHO one of the strong points of NeXTstep was that it is not going 
beyond Unix but integrating it. And the same is true (to a lesser 
extend, Finder argh) for MacOSX.

Greets,
  Helge
--
http://docs.opengroupware.org/Members/helge
OpenGroupware.org