mh@wlbr.EATON.COM (Mike Hoegeman) (07/08/88)
In article <10250002@hpfclp.SDE.HP.COM> diamant@hpfclp.SDE.HP.COM (John Diamant) writes: >Until now, it was a simple choice: do you want a proprietary system which one >vendor is trying to foist on the world (NeWS) or a multi-vendor, open system, >which has been freely available from the very start (X). > Hmm... I don't know if i would call it foist. I think sun is bending over backwards to accomodate the X fans of the world w/ the merged NeWS/X11 server. I think sun's attitude seems to more "may the best one win". sun's policy on NeWS is pretty much like NFS , if you want a reference source kit you hand over your XXXX amount of dollars and you get it. >NeWS advantages over X: >high power imaging model (2D) using absolute dimensions, rather than pixels >non-rectangular windows >Postscript available Not just available . postscript it is completely and pretty much seamlessly an integral part of the server. the current plans for postscript under X just seem like some weird abortion in comparison. >Toolkits can be interpreted in the server, and thus substituted out from under > the application >mimimal traffic between client and server > >disadvantages of NeWS relative to X: >requires relatively powerful NeWS server -- a NeWS terminal will be more > expensive than an X terminal I don't know if this is so true either. X is no doubt more thrify in resources than NeWS but once you start using those monster X libraries (and you have to to get anywhere with it) I'm not so sure the total resource usage will be in X's favor. NeWS runs quit nicely on Macs and amigas and 386's. I don't really think anybody is going to bother running X or NeWS on anything smaller than those type of machines. > >programming process context switching and partitioning between client and > server is a PAIN for the progammer. Yeah, It's a pain sometimes. but I think It's well worth it just to get the postscript paradigm >programming in Postscript is a pain (of course, Sun provides a C translator > so this isn't that big a problem). I would'nt say that necessarily. In fact I know alot of people who really like postscript. I really did'nt like it much at first but now i think it's great! You really have to USE it for awhile to realize how powerful it is. This is , I think, is one of the major advantages of news over X. the imaging model is just so much more powerful and flexible. Plus it's INTERPRETED!! It's great to go in and whip up a prototype window or user interface gadget by just cranking up a postscript shell and having at it! While on the subject of PostScript , there's the implementation of objects and classes to consider too. This is a real boon when creating user interface gadgets. >The ability to have interpretive toolkits which can be swapped is useful, but >there are other ways to accomplish this in X as well (using dynamic loading, >for instance). >Probably the most significant difference between X and NeWS is the traffic >between client and server. First of all, Scheifler wrote a paper about why It's nice but I don't think anybody really goes for NeWS because of this. I don't. > >> One more note : I tend to think that those companies that >> offer both systems are in a much better position than those >> companies only offering X... How true.. -mike
jim@expo.lcs.mit.edu (Jim Fulton) (07/08/88)
> NeWS runs quit nicely on Macs and amigas and 386's. I don't > really think anybody is going to bother running X or NeWS on anything smaller > than those type of machines. It depends on whether or not you count the new window system terminals as smaller machines. Also, people are already using machines with 80286's to run window system servers (and have been for over 2 years). It might not be blindingly fast, but for organizations that have a heavy investment in PCs, it is a good solution. Jim Fulton MIT Laboratory for Computer Science
diamant@hpfclp.SDE.HP.COM (John Diamant) (07/12/88)
> Hmm... I don't know if i would call it foist. I think sun is bending > over backwards to accomodate the X fans of the world w/ the merged > NeWS/X11 server. Yes, the X11/NeWS server should satisfy everyone, but take a look at the history behind it. Also my comment was directed at the situation prior to having the X11/NeWS server. They really didn't have a lot of choice in the matter. Customers were demanding X for interoperability and the only way to get some of the vendors to look at NeWS was to provide it on the AT&T tape in the X11/NeWS merge (due to the large source licensing fee). Until it was clear that X11 was here to stay, Sun kept trying to make NeWS the standard and kill X. Only when it was clear that it wasn't going to work, did they join the bandwagon. > I think sun's attitude seems to more "may the best > one win". sun's policy on NeWS is pretty much like NFS , if you want > a reference source kit you hand over your XXXX amount of dollars and > you get it. Yes, that is true now. However, it is virtually impossible to implement either NFS or NeWS without the reference source. > >NeWS advantages over X: > >high power imaging model (2D) using absolute dimensions, rather than pixels > >non-rectangular windows > >Postscript available > > Not just available . postscript it is completely and pretty much seamlessly an > integral part of the server. the current plans for postscript under X just > seem like some weird abortion in comparison. Maybe you mean "aberration." Anyway, you're right that Postscript integration in NeWS is superior to the Postscript on X work. I didn't mean to imply that it wasn't. My separation of the imaging model from Postscript availability attempted to suggest that, but apparently I failed. > >disadvantages of NeWS relative to X: > >requires relatively powerful NeWS server -- a NeWS terminal will be more > > expensive than an X terminal > > I don't know if this is so true either. X is no doubt more thrify in resources > than NeWS but once you start using those monster X libraries (and you have > to to get anywhere with it) I'm not so sure the total resource usage will be > in X's favor. NeWS runs quit nicely on Macs and amigas and 386's. > I don't really think anybody is going to bother running X or NeWS on > anything smaller than those type of machines. I believe I have seen X product announcements for X on an IBM PC (8088). I'm pretty sure I've seen recommended configurations as low as 286 at least. I'm thinking in terms of dedicated X or NeWS terminals. The problem with NeWS terminals is that someone might try to run large portions of their program in the terminal, and it won't be able to handle it. It could probably handle reasonable client/server mixes, but there is nothing to prevent a NeWS program from attempting to run almost entirely in the server, which I'm sure would fail miserably on a NeWS terminal. > >programming process context switching and partitioning between client and > > server is a PAIN for the progammer. > Yeah, It's a pain sometimes. but I think It's well worth it just to get > the postscript paradigm Only if it saves you in software development cost elsewhere. If it makes developing the software harder, I don't think it's worth it. > > While on the subject of PostScript , there's the implementation of objects > and classes to consider too. This is a real boon when creating user interface > gadgets. Yes, that's true. Toolkits (on X) are object oriented too, of course. John Diamant Software Development Environments Hewlett-Packard Co. ARPA Internet: diamant@hpfclp.sde.hp.com Fort Collins, CO UUCP: {hplabs,hpfcla}!hpfclp!diamant
root@sbcs.sunysb.edu (root) (07/15/88)
In article <10250003@hpfclp.SDE.HP.COM>, diamant@hpfclp.SDE.HP.COM (John Diamant) writes: > to get some of the vendors to look at NeWS was to provide it on the AT&T > tape in the X11/NeWS merge (due to the large source licensing fee). Until it > was clear that X11 was here to stay, Sun kept trying to make NeWS the standard > and kill X. Only when it was clear that it wasn't going to work, did they join I've been to a couple rah-rah seminars for NeWS; its creators have shown considerable restraint towards bashing X. More so than I would have if I created a great product like NeWS. As for the "large source license fee", one hat I wear besides the one at Stony Brook is that of a small company. My partners and I were able to afford a NeWS license - I can't see why anyone else would have a problem. Especially HP :-). I've not spoken to anyone on the NeWS project that had the attitude that they're out to build NeWS to foist it on the world, kill X11, etc. It seems that they have their view of what a network window system should look like and they've been pursuing it. > the bandwagon. > > > I think sun's attitude seems to more "may the best > > one win". sun's policy on NeWS is pretty much like NFS , if you want > > a reference source kit you hand over your XXXX amount of dollars and > > you get it. > > Yes, that is true now. However, it is virtually impossible to implement either > NFS or NeWS without the reference source. > Untrue about NFS. We implemented an NFS client for the Amiga from the protocol spec. No problem at all. We will shortly introduce an NFS server product that was built, once again, from the publicly available specs. NeWS would be a tougher nut to crack, but I think it is doable. I've done my share of Sun bashing, but one area you really can't touch them on is they do make their technology available, and for pretty reasonable fees. > > I believe I have seen X product announcements for X on an IBM PC (8088). I'm > pretty sure I've seen recommended configurations as low as 286 at least. I'm > thinking in terms of dedicated X or NeWS terminals. The problem with NeWS > terminals is that someone might try to run large portions of their program > in the terminal, and it won't be able to handle it. It could probably handle > reasonable client/server mixes, but there is nothing to prevent a NeWS program > from attempting to run almost entirely in the server, which I'm sure would > fail miserably on a NeWS terminal. NeWS running on the Amiga isn't terribly fast compared to even NeWS on a 3/50. The important point to remember here is that a small machine running X or NeWS is delivering price performance rather than graphics performance. As to whether one needs less machine to run X, here is a suprising result: The alpha binary for one X11 server on the Amiga is actually 40Kbytes larger than our current NeWS binary. The performance seems to be similar. Both ports require similar machine resources to run, eg ~2-3 mBytes ram, ~10 mBytes fonts, etc. Both ports are roughly at the same stage of maturity. Draw your own conclusions - the one I like is that it is just too early to say that either X or NeWS inherently require less machine than the other. > John Diamant > Software Development Environments > Hewlett-Packard Co. ARPA Internet: diamant@hpfclp.sde.hp.com > Fort Collins, CO UUCP: {hplabs,hpfcla}!hpfclp!diamant Let me just say that despite my obvious bias towards NeWS, I most certainly not anti X. All of this X vs NeWS vs etc nervous energy ought to be directed towards standardizing user interfaces so that we can get on with getting reasonable computers into mass markets. In the end, does the customer really care whether it is PostScript, X RPC, etc on the wire? Of course not. Rick Spanbauer SUNY/Stony Brook
weiser.pa@XEROX.COM (07/24/88)
"Yes, that is true now. However, it is virtually impossible to implement either NFS or NeWS without the reference source." There are counterexamples to this. There are at least two no-cost implementations of Postscript, both capable of writing to screens, done with using the reference source. And of course many people have done NFS from the specs. We here have done one of each (although ours are not publically available). -mark