On Wed, 2003-03-12 at 23:56, Gerald Henriksen wrote: > Key features of WS apparently are JRE 1.3.1, Gnome 1.4, XFree 4.1.0. > The markets WS is meant for include S/W development and ISV > applications including the Oil/Gas field. > > So lets look at S/W development. A significant percent of S/W > development is going to be GUI apps. Who is going to be developing > Gnome 1.4 applications for the next 12-18 months? Almost any > commercial development being done today for Gnome (remember the price > of WS, only companies will be buying this for the most part) will be > using Gnome 2 both for the better font support as well all the other > things Gnome 2 brings to the table. > > Furthermore most Java development should have shifted to Java 1.4.1 by > now. > > Applications in the Oil/Gas field are likely to be graphical, as are a > lot of things that a WS (and not desktop) operating system is aimed > at. Yet we have an older version of XFree that apparently has limited > 3D support (remember, 3D != games for many people).
I think the key to understanding this move is to remember that companies (to whom this product is being marketed) don't tend to live on the bleeding edge. When it comes to writing software, it's about writing software for the majority of your target demographic. The Gnome 2 stuff is still getting the bugs worked out of it. (I like the look and feel of Gnome better than KDE, but the inconsistency of the look and feel between the apps that make up "Gnome 2" is painful.) It's akin to writing Windows software. You can't write to the Windows XP "Luna" interface exclusively; you most likely still have to support 98 and 2000 clients as well. The listed applications are nice, but I imagine that there are 2 applications for which this distro will be most used: 3D graphical engineering workstations, and 3D special effects workstations (for Hollywood). I'm currently working on a report at my workplace that shows if we can use Linux or Windows in place of our Unix workstations for engineering development. We use lots and lots of packages like NASTRAN, STAR-CD, Hypermesh, etc. In fact, we use something like 16 different, high-end finite element codes. Almost all of them have at least a solver that runs on Linux, and most of them do this to support Linux clusters. Note that a "cluster" here isn't a cluster in the RH AS server sense; it's a group of machines running free scheduling software to distribute the load. So using an EW-type distro would be ideal for this. It would also be great to use it as a front-end client for this type of work. Only about half the codes we use have a Linux client, but that means that certain people would be in good shape to make the switch, and use a $3000 dual-Athlon SCSI-based workstation that will blow the pants off the $25,000 dual J-class HP workstations we're currently buying. Yes, 3D support will be limited, but if you choose your hardware correctly, it can still be done. The biggest hindrance to implementing an EW-type distro will, of course, be Windows application compatibility. RedHat needs to be taking steps to eliminate this problem. I don't know what that might entail. Perhaps just including a Citrix client on the disc would be an answer. Perhaps it means including Crossover Office. Maybe it's a credit with Citrix or Codeweavers for a copy of their server product for every 10 licenses purchased. I don't know the best way to handle this, but it's something that I think most every "big company" -- again, to whom this is being marketed -- is going to have to deal with (including us). I'm sure there will be volume discounts, but you can't expect a company to buy a "PC" OS for $300, and not have them expect to be able to run Office on it. Perhaps there will be even more information forthcoming about this, as it seems to be a gaping hole in their announcement. dk -- Phoebe-list mailing list [EMAIL PROTECTED] https://listman.redhat.com/mailman/listinfo/phoebe-list
