My donation: Of course not every user can change software on machines in every hospital. Users can't do a thing. One of the highlights of a Unix (and of course, a Linux) management, is that it can be managed from one sigle center point, and easilly. Assuming, for the matter, that there are zero hardware failures (we'll get there soon), all you need to do is implement the system once, and manage it using minimal number of people - You could scale up (as I see it) to having a very small amount of experts, each in his field, regarding parts of the whole system - You could have a DB expert, a scripting expert, a testing expert (working with the rest towards new concepts, updates, changes, etc) and half a dozen Sysdamins for the whole system, all in all. After sufficient testing, one could hardly argue the stability of an open-source (free?) software based systems. I don't remember a link, and many details, but there was a whole city in the US, if I remember correctly, which was managed (police, offices, etc.) by 3 people, including the boss, with lots of time on their hands. The only drawback of not using a mainframe (and there's nothing we can do about it) is hardware problems. That implies that every hospital should have one, two, even three technicians. Since management can be done remotelly, and central based, you only need to keep these techies on the spot to: - maintain and fix hardware problems - "feel" the existing system - They will probably know first if there's any new problem - support individuals. Using central maintanace, personal usage profiles would be very stable, and could always be reset if the need arises (which should be quite rare).
For example - assume usage of one LDAP server in every hospital, containing data about net-home-dir, managed from one central point, and updated from this central point, as well as any and every other LDAP servers the Clalit has. What you get, after building the right scripts (it's being called "deploying") the system, is that your "users administrator" inserts a single user into LDAP, all LDAP servers update, a homedir, preferences preconfigured, is being openned on the correct "area" storage server, mail is being configured, and vualla! A computer is added into the hospital - a Network KS is being deployed (pre-existing), configuring the computer to be part of the area, activating a "register" request for the "Users admin", or net-admin, gets preconfigured according the the legitimate std. configuration in the area, during the install, and is waiting for the users to use, in no time (a proccess like that takes around an hour. If you use very long and complicated scripts, that is, else it can be less then 20 minutes). The only actual job was to connect the computer to the net, and boot it up using floppy, cd, or net card. The system knows what to do next. Cheapper on deployment, cheapper on management, cheapper to use. Deployment, of course, takes time. One should be capable enough to plan a good layout - extensible, scalable, and easy to maintain, to allow easy maintanace, and easy configuration. Oh, and you don't need to have your FSMOs. Etzion Bar-Noy Aka, Ez. On Tuesday 04 November 2003 10:47, Shachar Shemesh wrote: > Hi list, > > Just came back from a meeting with Gadi Gilon. For those who don't > remeber, he is the CIO of "Kupat Cholim Klalit". He stumbled upon the > last time his name was mentioned on this list (thread starting at > http://plasma-gate.weizmann.ac.il/Linux/maillists/03/09/msg00296.html), > and wanted to talk. > > I'm BCCing him on this email, so he can choose to participate actively > in this discussion (or just correct me). > > The discussion stayed, almost exclusively, on the theoretical, > ideological, front. If I understood correctly, his main point is this: > "I can see ideological/social reasons for writing/using free software, > and I see financial ones. If I try to adopt the ideological reasons > within my organization, it will never work. I cannot let every user of a > machine in every hospital change their own software. I cannot expect to > have the social contract's benifits when aquiring the software, yet not > pass the same benifits onwards. I must therefor reject the social > reasons for adoping free software". > > Please don't start flame wars saying "but there are also financial > reasons for adopting open source". He is not rejecting this possibility. > It has not come up due to lack of time. > > Now, I tried to point the practical reasons behind the social contract, > and his response rather suprised me. Basically, he has contracts with > all of his software vendors that gives him full access to the source > code in case the company goes under. His basic premesis was "I can get > competition over support in proprietary software too - Clalit did it in > the past already". > > I tried to point out that this is actually means that he has forced the > vendors to turn their model, when dealing with him, into a free software > one. He acknoledged the possible truthfulness of this statement. I read > that as "the free software model is so much suprior, that I am actually > forcing closed source companies to adhere to it". I guess you may read > that to mean that the free software advantages are not as important to > him, representing a huge organization, as it is to SMBs. I guess had the > quotes in the paper said "I don't see the advantages of free software to > organizations of my caliber", things would have been understood > differently by us too. > > Like I said before - the financial aspects of free software were not > raised at all. > > One last point - he said he is willing to talk to us further. If he > doesn't wish to join this discussion (and even if he does), I was > thinking of dedicating a Telux meeting to that end. What do you think? > > Shachar ================================================================= To unsubscribe, send mail to [EMAIL PROTECTED] with the word "unsubscribe" in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
