eric@tcgould.tn.cornell.edu (Eric Fielding) (03/22/89)
Reply-to: eric@tcgould.tn.cornell.edu (Eric Fielding) I am not sure whether to post this on the Ultrix or VMS newsgroup, so I hope you will excuse my posting to both. We are considering the purchase of a DECstation 3100 to add a fast Ultrix machine to our facility. The question is how well we will be able to make use of our existing Local Area VAXcluster of VMS workstations and file servers. We hear that the VMS/Ultrix Connection allows a VMS machine to serve Ultrix machines via NFS. Is this a reality or a plan? Does it make sense to buy a diskless or nearly diskless DECstation 3100 to run off a VAXstation 3200 file server? While I am asking questions, how much memory is reasonable for an Ultrix machine? Our VAXstation 3200 with 8 MB is noticeably lacking in memory. Finally what are people's feelings about the DS 3100? Is this the wave of the future or another DEC misstep like the DEC Pro or Rainbow? ++Eric Fielding eric@geology.tn.cornell.edu (I'll summarize if there is interest)
rsp@decvax.dec.com (Ricky Palmer - (603)881-0370 - ZK3-3/T74) (04/12/89)
------- You can't go wrong with the DS3100. My suggestion is to get it with 12-24 MB of memory, the local paging/swapping disks and run diskless. If you can afford it, running it off of its own RZ55 is the best. The Ultrix/Connection product is also worth its while since you can then serve up your VMS databases over NFS. Enjoy! I am. Ricky Palmer --- Ricky Palmer Ultrix Advanced Development --- Digital Equipment Corporation --- Nashua, New Hampshire --- ... One of the fathers of PMAX ...
alex@wolf.umbc.edu (Alex Crain) (04/13/89)
In article <4706@decvax.dec.com>, rsp@decvax.dec.com (Ricky Palmer - (603)881-0370 - ZK3-3/T74) writes: > You can't go wrong with the DS3100. Sure you can. Just get it delivered with 8meg and watch it swap itself to death. My configuration = 8Mb + 32Mb swap over the ether. X server, a few utilities (nothing large) and no decwindows. I can't compile two C programs simultaniously without getiting "out of swap space" errors, the linker is *slower* then the vax linker on a 2000. I'm trying to port KCL, and I've *crashed* the machine twice simply by overloading it! At no time have I attempted to do something that would not work on my 2000. My suggestion is to get it with > 12-24 MB of memory, the local paging/swapping disks and run diskless. My boss wants to use one ofthese things as a server for 20 or so users coming in off serial lines. How much memory do we need? 100Mb? 150? I was figuring on loading it up with 24Mb and adding a Maxtor as a swapper, and then directing user complaints to someone else. will this work? > --- Ricky Palmer Ultrix Advanced Development > --- Digital Equipment Corporation > --- Nashua, New Hampshire > --- ... One of the fathers of PMAX ... Are you sure this thing cooked for the full 9 months? how soon do we get a less buggy software release? :alex Alex Crain Systems Programmer alex@umbc3.umbc.edu Univ Md Baltimore County umbc3.umbc.edu!nerwin!alex
rsp@decvax.dec.com (Ricky Palmer - (603)881-0370 - ZK3-3/T74) (04/14/89)
-------
You might want to take a look at what is causing your problems:
If your machine is swapping to the ethernet frequently it is very
probably because it just plain doesn't have enough memory for all that you
want to run. Two solutions: buy more memory and/or install a
local disk(s) for all the swap activity. PMAX has some of the best
performance I have seen for swapping to the RZ55 or even the slower
RZ23. I spent a good deal of time last summer optimizing large request
performance in the SCSI code (overlapped I/O) so I know this works well.
I am surprised that you can "crash" the machine just by swapping. I would
like to see this in person since it should not happen. If it is happening
we should know about it. Other people like Jim Gettys have run
extensive swapping tests all with glowing successes! Also, when it swaps
over ethernet what is it swapping to? Is it going to your 2000? Argh!
Also, if you are porting code AND running DECwindows you definitely should
have more memory. DECwindows is a memory PIG. Get more memory is not
only good advice but necessary. Swapping should only
occur when necessary not all the time!!!!
As for the linker, I am not surprised if it is slower. All RISC systems
suffer from this same problem. The output is typically 30 percent "larger"
than equivalent CISC code.
>>>>12-24 MB of memory.... In my article I didn't know I was talking about servers!
The machine makes an excellent server.
--- Ricky Palmer Ultrix Advanced Development
--- Digital Equipment Corporation
--- Nashua, New Hampshire
--- ... One of the fathers of PMAX ...alex@wolf.umbc.edu (Alex Crain) (04/15/89)
In article <4716@decvax.dec.com>, rsp@decvax.dec.com (Ricky Palmer - (603)881-0370 - ZK3-3/T74) writes: > You might want to take a look at what is causing your problems: > > If your machine is swapping to the ethernet frequently it is very > probably because it just plain doesn't have enough memory for all that you > want to run. No kidding? Like I said, my machine has (only?) 8Mb. It swaps over the ether to another 3100 with an RZ23, 32 meg allocated allocated for swap. In a word, performance is abysmal with this configuration. As best I can figure, this machine should have a minimum of 16 Mb, and a local swap disk, probably an RZ55. Add those things, and you've got a dynomite single-user workstation. It probably wouldn't crash so much, either. But thats not what DEC is selling. I get the impression from the announcements and demos that this thing is supposed to work out of the box, and it doesn't. You probably should have 24Mb to run this thing diskless and use it as anything other then a glorified terminal. > I am surprised that you can "crash" the machine just by swapping. I would > like to see this in person since it should not happen. I'm not sure that this is crashing *because* of swapping, but it does crash under stress, and I can make it crash If I beat on it hard enough. I believe that this should not happen, but I don't care enough to go digging through the kernal to fix a supported product. I reported it and was informed that it didn't happen, so I must have dreamed it. > Also, if you are porting code AND running DECwindows you definitely should > have more memory. DECwindows is a memory PIG. I'm not running DEC windows. Here's a scenerio. I'm running a normal system (kernal, syslog, portmap, ypbind, biod, sendmail, xdm, Xmfb, update, cron, inetd, rwhod, lpd, elcsdm) and some user processes (uwm, xclock, xbiff, xload, xterm, xterm). Now for the heavy stuff. If I try to link KCL (a 2.4 Mb binary) while running emacs, I will probably run out of swap space. If I try to make something else (repeatedly exec cc), The machine will probably crash. I'm not impressed. :alex Alex Crain Systems Programmer alex@umbc3.umbc.edu Univ Md Baltimore County umbc3.umbc.edu!nerwin!alex
eric@pprg.unm.edu (Eric Engquist [CoE]) (04/15/89)
I have a question of either Jim Gettys or Ricky Palmer. (or for that matter any one else out there) The DECstation 3100 is extremely fast. However, has anyone tried to use the server model of it to support a large user base? In particular we have some old 780's with a user population of 1000. Usually about 20 or so are logged in via dumb terminals doing simply tasks, such as vi'ing, fortran, C, pascal compiles. (e.g. work a freshman - junior level engineering student would do) If I have a decstation 3100 with 24MB of memory and 1.5Gb of disk, how many users could I support concurrently on this machine? rough estimates are fine. My personal guess would be 2-3 times what the 11/780 could. Also, on a related subject. Suppose I want to support some X-windowing terminals on this system. The users are doing the same type of things. How may X window terminals could I reasonally support per a fully loaded DECstation? Also, if I build a lab with 2 Decstation 3100's and 20 xwindow terminals, should I hide that ethernet behind a Lan Bridge. (i.e. is the network traffic going to be unreasonable. I will probably find these answers out myself shortly, but any insight can be extremely helpful. Thanks, Eric Engquist Univ. of New Mexico College of Engineering eric@sybil.unm.edu (505)277-5501 ---Oh shit! You did it just like I told you to...
avolio@decuac.dec.com (Frederick M. Avolio) (04/15/89)
This is as good a time as any to remind people of something important. I personally don't care if you flame Digital or not or if you like or hate Digital. (Okay... that's a lie. I *do* care a lot!) But let me suggest that while a news group is a very useful thing for sharing problems, triumphs, questions, and answers, it does not replace or preclude other avenues for problem resolution. Bugs should *always* be submitted via SPRs. Problems with the documentation or requests for enhancements also. In addition, Digital documentation usually has mail-in cards in the back for comments on documentation and manuals. You have any idea how many the doc group usually gets? Under 10 is probably close to the mark for each SW release. In addition, if you deal directly with Digital, you should make your sales rep. aware of any problems. Yes, I have heard stories of unresponsive sales reps, but as they *need* your business to stay in business and that usually implies that they *want* you to be happy -- and obviously (isn't it?) this goes for *any* one selling a product and counting on repeat sales -- they are usually responsive and want to help. Sales reps have at their disposal -- often/usually/they'd better -- software support. That is, for example, one of my jobs. Supporting the people who support sales. If I am hearing about a problem that a *local* customer is having -- UMBC, for example -- *first* via the USENET, then there is a communication problem. Finally, some of you pay big bucks for Software Support from Digital -- or whatever company for that matter. You're throwing money away if you don't use what you are paying for. And if you don't think you are getting your money's worth you should complain loud, long, and clear. All I'm saying is, use all avenues available to you. I bet we can solve or resolve a problem. I'd love to try. Fred
aad@stpstn.UUCP (Anthony A. Datri) (04/17/89)
In article <2661@decuac.DEC.COM> avolio@decuac.dec.com (Frederick M. Avolio) writes: >In addition, if you deal directly with Digital, you should make your >sales rep. aware of any problems. Yes, I have heard stories of unresponsive >sales reps, but as they *need* your business to stay in business and that >usually implies that they *want* you to be happy My sales rep has done the following: o told me that VMS VWS tapes will work under Ultrix (as UWS) o been incredibly unresponsive o failed to be able to tell me what the items on my software support contract MEAN. At this point, I've been through 2 referrals, and everyone says they'll get back to me, and they haven't. I give up. I can't even get an answer to a simple question like: If DECwindows aka UWS is packaged with Ultrix, and DECwindows obsoleting VWS is packaged with VMS, why are there lines for them on my contract? Why did I get a box with the UWS part number that included a whole duplicate set of Ultrix manuals? (it also kills me to pay $170 a month to get make and RCS, but that's another issue) -- @disclaimer(Any concepts or opinions above are entirely mine, not those of my employer, my GIGI, my VT05, or my 11/34) beak is@>beak is not Anthony A. Datri @SysAdmin(Stepstone Corporation) aad@stepstone.com stpstn!aad
bph@buengc.BU.EDU (Blair P. Houghton) (04/17/89)
In article <2661@decuac.DEC.COM> avolio@decuac.dec.com (Frederick M. Avolio) writes: >This is as good a time as any to remind people of something important. [...] >If I am hearing about a problem that a *local* customer is having >-- UMBC, for example -- *first* via the USENET, >then there is a communication problem. Aha! Make the mental leap. If there is a communication problem, but it appears that the communication is taking a different route from the one designated, which therefore implies that that route is a path of less resistance, then it seems the one looking for the message is looking the hard way... what I mean is, why don't DEC make known that they can be reached via Internet? Most of netland do most of our human communication this way. It's easy to ask the world about problems because we're used to seeing others' problems and wondering if we can help. What i mean is, Did you refer the person's problem to the proper DEC department, or did you just refer the person? There's a huge difference, and anyone who has either dealt with or abetted any sort of bureaucracy knows the difference. If it's easier to get a back-alley source-code fix than to fill out the forms, then I pray for the customer. >All I'm saying is, use all avenues available to you. I bet we can solve or >resolve a problem. I'd love to try. The only thing stopping you from doing it now is that you are _waiting_ to be asked personally. --Blair "Being a semi-professional buttinski has its rewards."
alex@wolf.umbc.edu (Alex Crain) (04/18/89)
Ahem, well. I feel that I own an apology to the DEC developers who read and respond to this newsgroup, regarding my comments on the 3100. First, a clarification: My opinions are my own, anyone foolish enough to agree with me deserves what they get. My comments about the 3100 were poorly aimed, and I'm sorry. I search of the lowest price, I think that DEC is shipping the machine in a smaller then useable configuration, a practice which introduces hidden costs to the customer (more memory/disk), and I find this irresponsible. There are also software bugs, but they are to be expected, and would probably be tolerable if the machine was correctly configured. In any case, none of this is the fault of the designers, and to you, I apologise. In fact, I think that compliments are in order for the degree to which you were able to preserve a VAX-ultrix like environment on an alien architecture. As far as my bitching to the net goes, thats a product of two things: (1) little response to my previous complaints, which may be dependent on my department as much as DEC. (2) experiance with DEC support that is so bad that even I have trouble believing it, and I was there. I've been instructed by my boss to use DEC support for problems, and will do so under extream duress, but otherwise they can all burn in hell. Conversely, My understanding is that the DEC sales-rep has been very responsive to my net complaints (I don't deal with them directly) and various efforts are being made to resolve my problems. Based on the speed at which things have happened, I think things will work out without furthur issue. :alex Alex Crain Systems Programmer alex@umbc3.umbc.edu Univ Md Baltimore County umbc3.umbc.edu!nerwin!alex
davis@karl.ucar.edu (Glenn P. Davis) (04/19/89)
Being able to email bug reports, SPR's, and or software support questions would be a BIG plus. Who want to hang on the phone or put a form in a typewriter? I mean, Digital sells on the idea of corporate networking... But what about internetworking? BTW, I got a call yesterday from I guy inside DEC (bellevue) who wanted a copy of some of our ftp'able software. The only way he could figure to get it was for us to send a tape... Glenn P. Davis UCAR / Unidata PO Box 3000 1685 38th St. Boulder, CO 80307-3000 Boulder, CO 80301 (303) 497 8643
brian@ms.uky.edu (Brian Sturgill) (04/19/89)
In article <3021@ncar.ucar.edu>, davis@karl.ucar.edu (Glenn P. Davis) writes: ... > BTW, I got a call yesterday from I guy inside DEC (bellevue) who wanted a > copy of some of our ftp'able software. The only way he could figure to > get it was for us to send a tape... ... If you think that is bad try getting the "Customer Support" people to send you e-mail.... Brian ----------- Brian Sturgill System Manager University of Kentucky Departments of Mathematical Sciences {uunet,rutgers}!ukma!brian brian@ms.uky.edu brian@UKMA.BITNET
treese@crltrx.crl.dec.com (Win Treese) (04/20/89)
In article <8237@boulder.Colorado.EDU> bobk@boulder.Colorado.EDU (Bob Kinne) writes: >I am having some problems installing ULTRIX on a DEC RISC processor. >The DS3100 has a 100MB winchester (too small, I know). I am trying >to install from a SCSI TK-50. After typing the boot command, the >tape runs for about 5 minutes, then I get the beginning of the >hardware listing. It always lists the CPU and FPU, sometimes also >lists the display, and then produces two exception messages. The >second is always [...more info deleted...] I'd have to check to be sure, but I believe that the 100MB disk is not supported as a system disk. It's certainly not big enough to put a reasonable Ultrix system on if you've left a reasonable amount of swap space. The exceptions are admittedly a poor way to fail if it's simply the software not wanting to go onto the disk. Of course, it might still be hardware... Win Treese Cambridge Research Lab treese@crl.dec.com Digital Equipment Corp.
steve@fnord.umiacs.umd.edu (Steven D. Miller) (04/20/89)
You're running a DS3100 diskless? In most cases, that's like taking a good runner and cutting his legs off, and then expecting him still to run fast... (Caveat: I don't even like running Sun-3/50s diskless, and let's face it, a 3/50 isn't in the same class as a DS3100.) I don't think that DEC is wrong for selling diskless 8MB DS3100 systems. There might actually be some people Out There who can use them -- I can think of a secretary down the hall who would probably be thrilled to have one. The onus is on the purchaser to make sure that the system being purchased will fulfill all the right requirements. That's why I have mostly 12MB (or more) DS3100s, and that's why I have at least one RZ55 on each one. (Sun will happily sell you a diskless Sun-4 of any variety, but would you buy a diskless SPARCstation 390?) I did notice that the performance difference between NFS mounting one's binaries and running them off the local RZ55 is substantial enough to feel. If you've got the older RZ55s without the readahead cache, you're not seeing more than 400K/sec off the disk, and then you lose some of that through NFS; it's no wonder. I expect to see a big improvement in with the later RZ55s (the ones that run more like 800K/sec), even across NFS. (To check your rev level on your RZ55s, do a 'scsi pb' at the monitor prompt, and if the rev level comes back as 700 or less, join me in hoping for a FCO...) I agree that the machine shouldn't crash under serious pounding, though. It should just slow down a lot. -Steve P.S.: I did a 'dd if=dxpsview of=/dev/null bs=8k', for the case where the dxpsview binary (1368064 bytes) was on the local RZ55, and where it was on a NFS-mounted filesystem. The times I got were as follows: NFS: 6 seconds (228K/sec) 7 seconds (195K/sec) 4 seconds (342K/sec, prolly out of buffer cache) average: 5.67 sec, 241K/sec RZ55: 4 seconds (342K/sec) consistently for three tries These numbers are totally off the cuff, but are, I think, somewhat indicative; they indicate around a 40% speed loss for going to NFS. So, if we get 800K/sec out of a RZ55, we'd maybe see 480K/sec off the same disk via NFS. This ignores a lot of other factors, so maybe we won't see that, either; at least it's a guess, that someone with a newer RZ55 could go test out. Mumble, mumble, mumble. Spoken: Steve Miller Domain: steve@mimsy.umd.edu UUCP: uunet!mimsy!steve Phone: +1-301-454-1808 USPS: UMIACS, Univ. of Maryland, College Park, MD 20742 Spoken: Steve Miller Domain: steve@mimsy.umd.edu UUCP: uunet!mimsy!steve Phone: +1-301-454-1808 USPS: UMIACS, Univ. of Maryland, College Park, MD 20742
cks@white.toronto.edu (Chris Siebenmann) (04/21/89)
In article <17013@mimsy.UUCP> chris@mimsy.UUCP (Chris Torek) writes: ... | SPR.... | Incidentally, the `-C' option does not work in xterm either. (It is | listed in the usage error xterm produces, and it is in the manual.) You can get almost the same functionality with 'cat -u /dev/xcons' in a window; you have to remember to kill the cat before killing X11, or your console appears to hang. Some quick hack-work on xterm could probably bring back the -C flag fairly easily. This trick works on X11R2 and X11R3, under Ultrix 2.2 and 3.0 at least and you don't seem to have to install the UWS software first, either. It is possible that the UWS xcons program will also work, assuming it does the same thing. I haven't tested it, since we don't have UWS installed. [trick originally from Stu Grossman <grossman@polya.Stanford.EDU>, in response to my querry here.] [and why do people give usefull articles distribution usa? there's Ultrix hacks up here in Canada too, you know...] -- "I shall clasp my hands together and bow to the corners of the world." Number Ten Ox, "Bridge of Birds" Chris Siebenmann ...!utgpu!{ncrcan,ontmoh!moore}!ziebmef!cks cks@white.toronto.edu or ...!utgpu!{,csri!}cks
awpsys@ultb.UUCP (Andrew W. Potter) (04/21/89)
In article <11537@s.ms.uky.edu> brian@ms.uky.edu (Brian Sturgill) writes: >> copy of some of our ftp'able software. The only way he could figure to >> get it was for us to send a tape... >... > >If you think that is bad try getting the "Customer Support" people to >send you e-mail.... > We communicate via e-mail with our local sales Rep. The mail gets delivered usually within minutes. The path is: (get this) Our Allin1 => DEC message-router => PMDF => Jnet => Bitnet => internet => DEC Easynet => DEC message router => DEC ALLin1. Needless to say the return addresses are mutilated beyond recognition but it get there and our DEC Salesfolk are able to originate messages back. The mail does 2 continental hops before delivery. - Andy -- Andrew W. Potter Email: awpsys@ritvax.BITNET Systems Programmer awp8101%ritcv@cs.rit.edu Information Systems and Computing Rochester Institute of Technology, Rochester NY, 14623 (716) 475-6994
brian@ms.uky.edu (Brian Sturgill) (04/22/89)
In article <716@ultb.UUCP>, awpsys@ultb.UUCP (Andrew W. Potter) writes: > In article <11537@s.ms.uky.edu> brian@ms.uky.edu (Brian Sturgill) writes: > >> copy of some of our ftp'able software. The only way he could figure to > >> get it was for us to send a tape... > >... > > > >If you think that is bad try getting the "Customer Support" people to > >send you e-mail.... > > > We communicate via e-mail with our local sales Rep. The mail gets > delivered usually within minutes. The path is: (get this) .... Several people have told me this... I guess they just don't like me there? Brian ---------- Brian Sturgill System Manager University of Kentucky Departments of Mathematical Sciences {uunet,rutgers}!ukma!brian brian@ms.uky.edu brian@UKMA.BITNET
rsp@decvax.dec.com (Ricky Palmer - (603)881-0370 - ZK3-3/T74) (04/22/89)
------- No, this is not due to the autoconfiguration code at all. It is due to the system not being in the proper state to begin with. Chris' comments about init'ing are correct to clear this up. Ricky --- Ricky Palmer Ultrix Advanced Development --- Digital Equipment Corporation --- Nashua, New Hampshire --- ... One of the fathers of PMAX ...
rsp@decvax.dec.com (Ricky Palmer - (603)881-0370 - ZK3-3/T74) (04/22/89)
------- Actually, Steve, your 480KB/sec number/interpolation is not far off the mark. Ricky --- Ricky Palmer Ultrix Advanced Development --- Digital Equipment Corporation --- Nashua, New Hampshire --- ... One of the fathers of PMAX ...
grr@cbmvax.UUCP (George Robbins) (04/22/89)
In article <17023@mimsy.UUCP> steve@fnord.umiacs.umd.edu (Steven D. Miller) writes: > > I don't think that DEC is wrong for selling diskless 8MB DS3100 systems. > There might actually be some people Out There who can use them -- I can > think of a secretary down the hall who would probably be thrilled to have > one. The onus is on the purchaser to make sure that the system being > purchased will fulfill all the right requirements. That's why I have mostly > 12MB (or more) DS3100s, and that's why I have at least one RZ55 on each one. > (Sun will happily sell you a diskless Sun-4 of any variety, but would you > buy a diskless SPARCstation 390?) Yes, but the onus is also on DEC and it's salespersons to recommend appropriate configurations to fit their best understanding of the customers needs, and also to avoid misrepresentations, either explicit or implicit, that a hardware configuration with known limitations will satisfy the customers requirements, without clearly presenting the tradeoffs. Hopefully they are doing this, as DEC has no obvious need of "wadda ya mean, you wanted both an engine *and* tires?" sales tactics... Personally I feel diskless work stations are a little bit like cars without engines, sure you can get around as long as there is somebody to give you a tow, but you'd be an awful lot happier if you had bought the conventional configuration. -- George Robbins - now working for, uucp: {uunet|pyramid|rutgers}!cbmvax!grr but no way officially representing arpa: cbmvax!grr@uunet.uu.net Commodore, Engineering Department fone: 215-431-9255 (only by moonlite)
russes@ryn.DEC.COM (Bob Russes) (04/22/89)
In article <132@crltrx.crl.dec.com> treese@crl.dec.com (Win Treese) writes: > >I'd have to check to be sure, but I believe that the 100MB disk is not >supported as a system disk. It's certainly not big enough to put a It's not supported. If I remember correctly, if you try to install Ultrix onto the 100Mb disk, the install procedure will tell you that the 100Mb isn't supported as a system disk. From what I've seen here, two 100Mb internals isn't enough... Bob
jmg@cernvax.UUCP (jmg) (05/05/89)
In article <2661@decuac.DEC.COM> avolio@decuac.dec.com (Frederick M. Avolio) writes: **** deleted good suggestions **** >Finally, some of you pay big bucks for Software Support from Digital -- or >whatever company for that matter. You're throwing money away if you don't >use what you are paying for. And if you don't think you are getting your >money's worth you should complain loud, long, and clear. > >All I'm saying is, use all avenues available to you. I bet we can solve or >resolve a problem. I'd love to try. I am happy to hear that you can (re)solve our problems. OK, this is VMS, not Ultrix, but we have complained loud and long, paid big Swiss Francs, sent in SPRs, told our local sales/support BUT we are told that a stupid little problem with the RBMS command show forwarding known physical addresses when on a big network, will not be fixed until some future major upgrade of hardware/software sometime on the not-so-near future. This is regrettably only one example: there are others. I suspect that, like most competent and conscientious support people, you do like fixing user problems. Only trouble is, you have $-counting administrators who stop you from doing so if they don't see a profit in it. However, the loss is hidden, being a drop in user satisfaction. ps. I tried the DS3100 and liked it, despite the inevitable bugs. -- _ _ o | __ | jmg@cernvax.uucp | | | | _ / \ _ __ _ __ _| jmg@cernvax.bitnet | | | | |_) /_) | __/_) | (___\ | (_/ | J. M. Gerard, Div. DD, CERN, | | |_|_| \_/\___ \__/ \___| (_|_| \_|_ 1211 Geneva 23, Switzerland