0 INFO-VAX	Mon, 28 Jan 2002	Volume 2002 : Issue 53      Contents: Re: bkjfn is an English word?  Re: bkjfn is an English word? P Re: BLISS pros and cons, was: Re: historical evidence of what went wrong  at DECD Re: Capellas redefines Industry Standard to mean Windows *and* Linux3 Re: comp.sys.dec, FINAL NOTICE of Newsgroup Removal  Re: Compaq financial conference  Re: Compaq financial conference  Re: Compaq financial conference  Re: Compaq financial conference  Re: Compaq financial conference  Re: Compaq financial conference  Re: Compaq financial conference  Re: Compaq financial conference  Re: Compaq financial conference  Re: Compaq financial conference  Re: Compaq financial conference  Cross Compiler to SUN/Solaris?" Re: Cross Compiler to SUN/Solaris?" Re: Cross Compiler to SUN/Solaris?" Re: Cross Compiler to SUN/Solaris?; Re: Gartner Group lies about VMS ... loses all credibility! 1 Re: historical evidence of what went wrong at DEC  HP 35470 DAT drive question ! Re: I/O Performance on Alpha 4100  Re: mount problems on cluster  Re: mount problems on cluster 3 Re: VMS system on the sourceforge.net Compile Farm? % Re: When to start a new c.o.v thread? 5 Re: writev() crashes with access violation on OpenVMS 0 Re: Younger recruits versus experienced veteransF Re: Younger recruits versus experienced veterans ( was Re:     The dem9 Re: [Q] Why is there the limit of 8 levels of directories   F ----------------------------------------------------------------------  % Date: Sun, 27 Jan 2002 23:33:37 -0500 ' From: Jim Becker <jbecker@ui.urban.org> & Subject: Re: bkjfn is an English word?, Message-ID: <3C54D4A1.83762043@ui.urban.org>   Bob Koehler wrote: > I >    "bkjfn"  OK, so what is this, and how did it get into the 50000 word 1 >    dictionary that set password checks against?   B Obviously, Bob, "bkjfn" means Bob Koehler's Just Fully Nonplussed!   --
 Jim Becker+ The Urban Institute (http://www.urban.org/) ' Encompass (http://www.encompassus.org/) . ESILUG (http://encompasserve.org/lugs/esilug/)   ------------------------------  % Date: Sun, 27 Jan 2002 23:34:22 -0500 ' From: Jim Becker <jbecker@ui.urban.org> & Subject: Re: bkjfn is an English word?, Message-ID: <3C54D4CE.BF76B841@ui.urban.org>   Bob Koehler wrote: > I >    "bkjfn"  OK, so what is this, and how did it get into the 50000 word 1 >    dictionary that set password checks against?   F Obviously, Bob, "bkjfn" means Bob Koehler's Just Fully Nonplussed! <g>   --
 Jim Becker+ The Urban Institute (http://www.urban.org/) ' Encompass (http://www.encompassus.org/) . ESILUG (http://encompasserve.org/lugs/esilug/)   ------------------------------   Date: 28 Jan 2002 01:26:58 GMT( From: peter@taronga.com (Peter da Silva)Y Subject: Re: BLISS pros and cons, was: Re: historical evidence of what went wrong  at DEC 2 Message-ID: <a329d2$1vaj$1@citadel.in.taronga.com>  = In article <5b86b9ee.0201220747.3f527e97@posting.google.com>, - Jeremy Barker <jeremybarker@email.com> wrote: @ >It's probably easier to translate C to BLISS than the reverse. G >Especially given that nothing in C can match the BLISS macro facility.   M That's what m4 was invented for. It works just fine as a C preprocessor, too.    --  @ Rev. Peter da Silva, ULC.	                                 WWFD?  F "Be conservative in what you generate, and liberal in what you accept" 	-- Matthew 10:16 (l.trans)    ------------------------------  # Date: Mon, 28 Jan 2002 02:47:58 GMT 1 From: "David J. Dachtera" <djesys.nospam@fsi.net> M Subject: Re: Capellas redefines Industry Standard to mean Windows *and* Linux ' Message-ID: <3C54BD05.C50DD544@fsi.net>   & paddy.o'brien@zzz.tg.nsw.gov.au wrote: > [snip]E > But didn't King Billie III, only a short while ago, claim that open F > source would not have existed without Microsoft?  I could never work' > out what he intended to mean by that.   C Just a thought, but maybe he meant that without resentment of M$ to < drive it, the open source movement would have less momentum.   --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  % Date: Mon, 28 Jan 2002 01:12:09 +0100 . From: Vladimir Woelfl <gandalf@palach.okri.hr>< Subject: Re: comp.sys.dec, FINAL NOTICE of Newsgroup RemovalE Message-ID: <Pine.BSF.4.33.0201280106200.77397-100000@palach.okri.hr>   & On Mon, 14 Jan 2002, Hans Vlems wrote:  ) > Since when does Kroatia run Usenet? ;-)   ' Kroatia is .hr, local name Hrvatska :-) H concerning the question... i really don't know, but there is very littleG activity on Croatian Usenet... I'm trying hard to change this, but what F could a man with DEC HiNote Ultra II laptop & Acer's i686 server do to change this?  G Why is anyway comp.sys.dec be removed? Or is that "admin@ateziod.ac.kr" % just somebody trying too play a joke?    >  > Hans > ? > Terry C. Shannon <terryshannon@mediaone.net> wrote in message 9 > news:L4m08.26078$JF.202537@rwcrnsc52.ops.asp.att.net...  > > B > > "Usenet Administration" <admin@ATEZTOD.AC.KR> wrote in message3 > > news:fgghpo7q5rfa7jmlyycqq1ygyujehcb@4ax.com... 8 > > > As many of you are aware, this newsgroup is slated6 > > > for removal on January 31st, 2002.  The vote was: > > > taken at the end of last year, and there was a large4 > > > marjority who wants to see this group removed. > > > G > > > If you are a regular reader, you need to make other arrangements.  > > M > > You're welcome to visit www.openvms.org and www.tru64.org for your Compaq * > > enterprise OS and product discussions. > >  > >  >  >  >    ------------------------------    Date: 27 Jan 2002 13:18:34 -0600+ From: young_r@encompasserve.org (Rob Young) ( Subject: Re: Compaq financial conference3 Message-ID: <+nQNvGhZbwp6@eisner.encompasserve.org>   n In article <hOp48.19633$vH6.1106472@bin6.nnrp.aus1.giganews.com>, "Bill Todd" <billtodd@metrocast.net> writes: > : > "Rob Young" <young_r@encompasserve.org> wrote in message/ > news:hFIIMOnEr+24@eisner.encompasserve.org... 8 >> In article <3C51E91C.AC58DFC0@videotron.ca>, JF Mezei  > <jfmezei@videotron.ca> writes: >> > "Main, Kerry" wrote: M >> >> It does this at the system/os level. Transactions are replicated to the J >> >> remote site for redundancy, so there is no read-write load balancingK >> >> across servers at multiple sites. However, at the server HW level, it A >> >> definately has features that OpenVMS / Alpha does not have.  >> >L >> > Note that the OS provides base services for automated failover, but theM >> > applications must take advantage of those services. Otherwise the tandem I >> > behaves like a VMS cluster with just a failover of a process with no  > assurance - >> > that current transactions are garanteed.  >> >I >> >> From a business perspective, OpenVMS (RTR) ensures a transaction is & >> >> never lost once it is committed. >> >K >> > The problem with RTR is that if the machine(s) that run RTR fail, then  > that >> > assurance is lost.  >> > >>C >> If you are using RTR, you are *most likely* spending the time to ? >> create the infrastructure (spending the money too) to ensure 7 >> hardware failure doesn't cause loss of transactions.  > $ > Let's try this once again, slowly: > H > The *definition* of a transaction *includes* the requirement that onceG > committed it cannot be 'lost', unless the storage on which its record L > resides is lost.  Thus the *only* issues involve loss of storage, and thisJ > can be guarded against by replicating said storage (typically mirroring,N > remote if site-disaster-tolerance is desired) - or, if quick recovery is notJ > required, by maintaining a remote recovery log which can be used to roll2 > forward an off-site backup copy of the database. > I > There are two other orthogonal issues being talked about here that have K > *nothing to do* with 'losing' transactions (save insofar as they guard in L > general against loss of persistent data).  One is high-availability, whichM > is required for systems that cannot tolerate down-time and usually involves L > the ability to continue processing on other hardware if failure of currentK > hardware is detected (but does nothing to protect the integrity of system B > operation if hardware fails without being detected).  Another isK > fault-tolerance, which is required to ensure (even if hardware fails in a N > manner that allows it to continue to operate but produces incorrect results)G > that the system produces correct results (at least if the software is M > correct) - and most fault-tolerant approaches result in higher availability L > as well, since they typically use (local, not remote) hardware replication( > to check results as they are produced. > G > Fault-tolerance can be achieved without special hardware support, but M > typically not with acceptable performance (since results must be checked at M > fairly fine grain - e.g., certainly prior to most disk writes).  VMS hasn't M > offered fault-tolerance since the special ftVAX hardware went away, but VMS = > does offer high availability via its clustering mechanisms.  > L > So, one more time:  there's nothing special about RTR in this respect - itL > just preserves the 'no loss after commit' guarantee that *any* transactionC > monitor must provide.  It may also offer transaction-level (as an I > alternative to system-level or storage-level) remote mirroring to guard H > against site-disaster data loss, but this is not fault-tolerance, just > high-availability. >   A 	All very well... yes RTR isn't the *only* mechanism that ensures = 	you don't lose transactions... but JF's assertion, following  	Kerry's comment was this:   Kerry says:   C From a business perspective, OpenVMS (RTR) ensures a transaction is   never lost once it is committed.   JF says:F The problem with RTR is that if the machine(s) that run RTR fail, then that assurance is lost.   E 	Implying that machine failure would be an issue (causing transaction  	loss) at which point G 	it becomes really very poor design if that is true.  Yes, I agree with E 	all that you say above in sorting this out and would add in Cantor's G 	case with split VMS clustering, business continued with an intervening F 	cluster transition as the downside(s) (of course as articles mention,B 	Europe and London were actually able to continue business whereby? 	local offices uptown Manhattan couldn't, due to comm failure).   B 	Other solutions (replay logs, remote nodes in hot standby) aren't? 	as attractive in my opinion.  But you do the best you can with  	what you have to work with.   				Rob    ------------------------------  % Date: Sun, 27 Jan 2002 14:30:32 -0500 % From: JF Mezei <jfmezei@videotron.ca> ( Subject: Re: Compaq financial conference, Message-ID: <3C545546.EC3C7C8E@videotron.ca>   Bill Todd wrote:L > That's a very good example of the value of transactions:  the behavior youI > experienced could *not* have occurred had it actually been handled as a M > correctly-distributed transaction using the standard industry definition of  > the term      E Is this the same type of "transaction" that vendors claim they can do K thousands per seconds, when bank transactions are in the 10-12 transactions L per second, and SWIFT transactions are measured in seconds per transaction ?  K The theoretical definition of a perfect transaction is meaningless when the : real world has implemented a totally different definition.    N > In particular, when a physical action (such as dispensing cash) is involved,M > if you use transactions correctly the *only* case in which there can be any E > question about whether or not that action occurred is when the node ! > *performing* that action fails    H The problem is that in the real world of interbank ATM access,  the cardH issuing bank does not control the remote ATM.  There are a lot of layersE (controllers, switching computers etc) between the ATM in Coober Pedy / Australia and your bank account in Pittsburgh.    M > using a small amount of redundant flash memory - such that when it recovers + > it can correctly ascertain what it did).    K But when the ATM recovers, it has no idea of whom it was talking to when it M failed so where would it send the message "I gave the cash" or "I didn't give K the cash" ????? This is especially true when the ATM controller fails and a J new controller is brough on-line, totally unaware of the ATM's transactionK context.  If the ATM sends a message "hey, I didn't give the cash", to whom J will that controller retransmit the message ? It won't know that the ATM'sG transactions was routed over multiple systems to some bank in a foreign H country and by that time, the bank in the foreign country will have longW closed the transactyion assuming success because it didn't receive a notice of failure.   G Remember that many systems were designed to be cost effective over X.25 ' networks that were charging per packet.     K This is why having Tandem as ATM switches solves a lot of problems  because N the context of a transaction is preserved even if a process/processor fails in a Tandem node.    I > As I said, you don't understand the definition.  Typical fault-tolerant I > systems have no software component whatsoever:  they rely upon hardware  > mechanisms  L Perhaps you should talk to the VMS engineers who worked on the FT-4000 faultL tolerant vax.  And you might want to takek to the BAS24 software developpers for the Tandem systems.    ------------------------------    Date: 27 Jan 2002 14:25:24 -0800( From: bob@instantwhip.com (Bob Ceculski)( Subject: Re: Compaq financial conference= Message-ID: <d7791aa1.0201271425.49085d65@posting.google.com>   Y JF Mezei <jfmezei@videotron.ca> wrote in message news:<3C52EE76.D4C4B51B@videotron.ca>...  > Bill Todd wrote:L > > By using transactions (and resubmission on hardware fail-over), which isP > > what the preceding discussion was about.  If you don't use transactions, the9 > > problem becomes a lot more general (and intractable).  > N > By your definition, it makes Windows fault tolerant. If the client has to beO > smart enough to resubmit the transaction when it fails, then you don't need a * > fault tolerant machine at the other end. > M > But it also has to do with data integrity.  If the failure occurs after the H > host has finalized the transaction, but before the client has received% > confirmation, then what do you do ?   K the key for me is to make sure that when I am posting to multiple rms index I files, and the interactive session line drops in the middle of posting, I K can continue the posting process uninterrupted which is why I want to spawn L my posting routines ... confirmation back to the client then is easy, if theK spawning process is still present, the dibol message manager will send back N either a complete or error status, otherwise, if it is not present (line drop)G I can have the spawned posting process send mail to the clients mailbox O indicating either success or error in posting the transactions ... so reporting > back to the client from a spawned process is not a problem ...   ------------------------------    Date: 27 Jan 2002 14:40:49 -0800( From: bob@instantwhip.com (Bob Ceculski)( Subject: Re: Compaq financial conference< Message-ID: <d7791aa1.0201271440.eb9bdcc@posting.google.com>  r "Bill Todd" <billtodd@metrocast.net> wrote in message news:<9SB48.43762$%b.2446633@bin4.nnrp.aus1.giganews.com>...7 > "Bob Ceculski" <bob@instantwhip.com> wrote in message 9 > news:d7791aa1.0201260602.32fe5e31@posting.google.com... 9 > > "Bill Todd" <billtodd@metrocast.net> wrote in message > > news:<9wu48.40254$%b.2050053@bin4.nnrp.aus1.giganews.com>... > K > > so how do you prevent, say a dibol executable posting to an rms file or  >  multiple I > > rms files from failing if a telnet or decnet over ip set host session  >  fails > > from a dsl line drop out?  > J > By using transactions (and resubmission on hardware fail-over), which isN > what the preceding discussion was about.  If you don't use transactions, the7 > problem becomes a lot more general (and intractable).  > 9 >   I am working on this very problem, and rms journaling L > > seems to be out because these accounting apps might chain to other execs >  andI > > pass data thru core common ... the only approach I can think of is to0 >  spawn theK > > posting processes and by using the dibol message manager have the batch 
 >  postingH > > exec send a complete message back to the other when its done or send
 >  detailsM > > of an error, if any ... any other ways to ensure a job completes during a  >  dsl > > line drop? > N > In other words, implement a somewhat-transaction-like mechanism on your own.N > But unless you create the equivalent of a distributed multi-phase Commit (orI > can make transaction duplication acceptable by making your transactionscD > idempotent), you can still have failure modes that cause problems. >  > - bill  L rms journaling, rtr, transactions, is fine, I understand that, but this is aK "single online interactive" transaction ... an rms record or records may be?M locked to prevent data corruption ... this is not batch, the integrity of thesO record being read in real time must be preserved until the change is posted ...DJ if an online posting failed, and the rms lock on the record or records wasJ released, you cannot have some app keep trying by rereading the previouslyL locked records and then post ... in the mean time, another user or app couldN have came along and modified the record, which makes that post then "corrupt"!K How does this relate to online interactive processing, which relies on the dO write which is based on the data on the screen at the time the record is locked-I completing before another subsequent read/write operation occurs?  Do you J understand I am talking about a single online interactive real time write?   ------------------------------  # Date: Sun, 27 Jan 2002 22:56:48 GMT * From: "Bill Todd" <billtodd@metrocast.net>( Subject: Re: Compaq financial conference@ Message-ID: <QA%48.68051$%b.4192322@bin4.nnrp.aus1.giganews.com>  5 "Bob Ceculski" <bob@instantwhip.com> wrote in messagey6 news:d7791aa1.0201271440.eb9bdcc@posting.google.com...   ....  L > rms journaling, rtr, transactions, is fine, I understand that, but this is aeJ > "single online interactive" transaction ... an rms record or records may beK > locked to prevent data corruption ... this is not batch, the integrity ofi theEF > record being read in real time must be preserved until the change is
 posted ...L > if an online posting failed, and the rms lock on the record or records wasL > released, you cannot have some app keep trying by rereading the previouslyH > locked records and then post ... in the mean time, another user or app couldtE > have came along and modified the record, which makes that post then 
 "corrupt"!L > How does this relate to online interactive processing, which relies on theJ > write which is based on the data on the screen at the time the record is lockedK > completing before another subsequent read/write operation occurs?  Do youeL > understand I am talking about a single online interactive real time write?  G What I understand is that you have no idea what a 'transaction' (in the>J industry-standard sense, e.g., as used in database software) is.  When youK have rectified this the questions you pose above will be answered (at leasthK when transactions are used, which was the subject under discussion when you  joined the conversation here).   - bill   ------------------------------  # Date: Sun, 27 Jan 2002 23:04:48 GMTa* From: "Bill Todd" <billtodd@metrocast.net>( Subject: Re: Compaq financial conferenceC Message-ID: <kI%48.152412$TC1.10774392@bin5.nnrp.aus1.giganews.com>h  2 "JF Mezei" <jfmezei@videotron.ca> wrote in message& news:3C545546.EC3C7C8E@videotron.ca... > Bill Todd wrote:J > > That's a very good example of the value of transactions:  the behavior you K > > experienced could *not* have occurred had it actually been handled as aeL > > correctly-distributed transaction using the standard industry definition of > > the term >t >eG > Is this the same type of "transaction" that vendors claim they can dot@ > thousands per seconds, when bank transactions are in the 10-12 transactionsL > per second, and SWIFT transactions are measured in seconds per transaction ?r >lI > The theoretical definition of a perfect transaction is meaningless when- the-< > real world has implemented a totally different definition.  E JF, you haven't a clue about what you're talking about and persist inCG spewing nonsense despite being told so.  Every supposed issue you raise0K below is handled by distributed transaction mechanisms (which are supportedn> among multiple, heterogeneous databases - even across separateJ organizations - that conform to X/Open standards in this area), and if youH ever stop shooting your mouth off and instead learn something about them? you'll understand that.  But I've run out of patience, so 'bye.8   - bill   > F > > In particular, when a physical action (such as dispensing cash) is	 involved,.K > > if you use transactions correctly the *only* case in which there can bea any<G > > question about whether or not that action occurred is when the node7" > > *performing* that action fails > J > The problem is that in the real world of interbank ATM access,  the cardJ > issuing bank does not control the remote ATM.  There are a lot of layersG > (controllers, switching computers etc) between the ATM in Coober Pedyo0 > Australia and your bank account in Pittsburgh. >eF > > using a small amount of redundant flash memory - such that when it recovers, > > it can correctly ascertain what it did). >eJ > But when the ATM recovers, it has no idea of whom it was talking to when itJ > failed so where would it send the message "I gave the cash" or "I didn't giveK > the cash" ????? This is especially true when the ATM controller fails and0 arL > new controller is brough on-line, totally unaware of the ATM's transactionH > context.  If the ATM sends a message "hey, I didn't give the cash", to whomL > will that controller retransmit the message ? It won't know that the ATM'sI > transactions was routed over multiple systems to some bank in a foreignsJ > country and by that time, the bank in the foreign country will have longF > closed the transactyion assuming success because it didn't receive a notice of failure. >eI > Remember that many systems were designed to be cost effective over X.25t) > networks that were charging per packet.o >e >iD > This is why having Tandem as ATM switches solves a lot of problems because,G > the context of a transaction is preserved even if a process/processor/ fails in > a Tandem node. >> >NK > > As I said, you don't understand the definition.  Typical fault-tolerantaK > > systems have no software component whatsoever:  they rely upon hardware> > > mechanisms > H > Perhaps you should talk to the VMS engineers who worked on the FT-4000 faultnB > tolerant vax.  And you might want to takek to the BAS24 software developpers( > for the Tandem systems.i   ------------------------------    Date: 27 Jan 2002 15:23:41 -0800( From: bob@instantwhip.com (Bob Ceculski)( Subject: Re: Compaq financial conference< Message-ID: <d7791aa1.0201271523.c79897f@posting.google.com>  Y JF Mezei <jfmezei@videotron.ca> wrote in message news:<3C5426E9.E9D49779@videotron.ca>...> > Jim Johnson wrote:I > > -- Transactions: These are *defined* as having ACID properties, which B > > include durability of the result.  Anything that doesn't offer: > > durability of the result isn't offering a transaction. > N > Here is an example of a "failure" I personally experienced in the early days+ > of interbank ATMs back in 1986 timeframe.  >   M like I just told Bill, this example is batch based processing, an atm reportssL to the user what it has on record at that request time, then batch processesJ are either to withdraw or deposit money ... the order doesn't really quiteL matter, because the final balance will still come out correct no matter whatJ the posting order of the tranasactions ... my situation is different ... IK inherited both a corporate and a route based accounting system in which theaM stores/writes are done for the majority of the time interactively ... online!dG When a user reads the rms record for update, the record is locked whicheJ preserves the data for the user up until the time they "write" to modify aI record and prevents another user or app coming along in the mean time andtK modifying that record which could lead to data corruption ... in the case adM telnet or set host interactive line drops, I must preserve the locked recordsnL and complete the write before another user/app comes along to read/write ...K that is why I can only see spawning the write process which will ensure therK the rms records stay locked until the write is complete then report back toaJ the user a complete or error status ... batch solutions will not work with: online interactive real time rms read/write conditions ...   ------------------------------  % Date: Sun, 27 Jan 2002 21:29:53 -0500a% From: JF Mezei <jfmezei@videotron.ca>l( Subject: Re: Compaq financial conference, Message-ID: <3C54B7A0.B5936423@videotron.ca>   Bill Todd wrote:G > JF, you haven't a clue about what you're talking about and persist insI > spewing nonsense despite being told so.  Every supposed issue you raiseeM > below is handled by distributed transaction mechanisms (which are supporteda@ > among multiple, heterogeneous databases - even across separateB > organizations - that conform to X/Open standards in this area),   K Just because there exist some standards that define what a "transaction" istM doesn't mean that those standards are used in banking systems that were setup L well before VMS or UNIX  could run on iron fast enough to even be considered for banking systems.  N Care to explain why, while magazines were talking about machines that could doS 1000 TPS, banks were struggling at the 10-15 banking transactions per second level?e  M Not every transaction in the real world fits the ideal "X-OPEN standard" thatoN you mentioned. A transaction is much more than something you do to a database.   ------------------------------    Date: 27 Jan 2002 19:02:59 -0800( From: bob@instantwhip.com (Bob Ceculski)( Subject: Re: Compaq financial conference= Message-ID: <d7791aa1.0201271902.52db46c1@posting.google.com>n  r "Bill Todd" <billtodd@metrocast.net> wrote in message news:<QA%48.68051$%b.4192322@bin4.nnrp.aus1.giganews.com>...7 > "Bob Ceculski" <bob@instantwhip.com> wrote in messagef8 > news:d7791aa1.0201271440.eb9bdcc@posting.google.com... >  > ...t > N > > rms journaling, rtr, transactions, is fine, I understand that, but this is >  aL > > "single online interactive" transaction ... an rms record or records may >  be-M > > locked to prevent data corruption ... this is not batch, the integrity of< >  theH > > record being read in real time must be preserved until the change is
 >  posted ...vN > > if an online posting failed, and the rms lock on the record or records wasN > > released, you cannot have some app keep trying by rereading the previouslyJ > > locked records and then post ... in the mean time, another user or app >  couldG > > have came along and modified the record, which makes that post thent
 >  "corrupt"!rN > > How does this relate to online interactive processing, which relies on theL > > write which is based on the data on the screen at the time the record is	 >  lockedcM > > completing before another subsequent read/write operation occurs?  Do youeN > > understand I am talking about a single online interactive real time write? > I > What I understand is that you have no idea what a 'transaction' (in theuL > industry-standard sense, e.g., as used in database software) is.  When youM > have rectified this the questions you pose above will be answered (at least*M > when transactions are used, which was the subject under discussion when youa  > joined the conversation here). >  > - bill  I although I have not used rdb or oracle "sub schema" db's, I understand in G general their logic in regards to updating ... however rms isn't rdb orP- oracle, so my question remains unanswered ...d   ------------------------------  # Date: Mon, 28 Jan 2002 04:18:39 GMTa* From: "Bill Todd" <billtodd@metrocast.net>( Subject: Re: Compaq financial conferenceC Message-ID: <zi458.154570$QB1.11532429@bin3.nnrp.aus1.giganews.com>a  2 "JF Mezei" <jfmezei@videotron.ca> wrote in message& news:3C54B7A0.B5936423@videotron.ca... > Bill Todd wrote:I > > JF, you haven't a clue about what you're talking about and persist infK > > spewing nonsense despite being told so.  Every supposed issue you raiseoE > > below is handled by distributed transaction mechanisms (which aret	 supporteduB > > among multiple, heterogeneous databases - even across separateC > > organizations - that conform to X/Open standards in this area),e >aJ > Just because there exist some standards that define what a "transaction" isI > doesn't mean that those standards are used in banking systems that werew setupaC > well before VMS or UNIX  could run on iron fast enough to even bei
 considered > for banking systems.  J So what?  We were talking about RTR and NonStop Tandem, which talk exactlyI the same kind of transactions I'm talking about (though IIRC RTR may havenE additional restrictions requiring that resource manager operations beeD idempotent so that they can be redone safely).  If some antedeluvianG predecessor didn't, then it simply didn't implement transactions in thedD now-standard sense, and that's in no way relevant to the discussion.   >sG > Care to explain why, while magazines were talking about machines that- could doG > 1000 TPS, banks were struggling at the 10-15 banking transactions perr
 second level?D  G Because even though the transactions share the same definition they cani+ differ radically in size and/or complexity.4   >BJ > Not every transaction in the real world fits the ideal "X-OPEN standard" thatF > you mentioned. A transaction is much more than something you do to a	 database.f  J In fact, every transaction in the real world *does* fit the definition I'mK using - because it *is* the industry definition.  Anything that doesn't fit L it is not a transaction in the technical sense.  Period.  And the definitionH is in no way restricted to conventional databases:  file systems such asC Transarc's Structured File System offer transaction support, and myfL impression is that even RMS can be used (possibly with some aid from anotherC component) as a 'resource manager' element within such conventionalm8 transactions under the control of a transaction monitor.  > You're a real idiot, JF:  just as you did in pressing your ownL misconceptions about the relationships among VMS, Cutler, and NT a while agoL despite being corrected by people acquainted with the facts, you steadfastlyF stick to your own ignorance in this area even when it's pointed out byL people who know what they're talking about.  Shut up and learn something for	 a change.o   - bill   ------------------------------  % Date: Mon, 28 Jan 2002 01:48:14 -0500i% From: JF Mezei <jfmezei@videotron.ca>h( Subject: Re: Compaq financial conference, Message-ID: <3C54F42B.BC53F1B1@videotron.ca>   Bill Todd wrote:L > So what?  We were talking about RTR and NonStop Tandem, which talk exactly2 > the same kind of transactions I'm talking about   J I am looking at the big picture. From an ATM in the middle of australia toH your bank mainframe in the USA.  Your bank's mainframe cannot drive thatM machine in Australia because it doesn't know what type of machine it is, what>D its keypad layout it is what software is loaded into it and onto itsM controller etc. It is the australian bank's infrastructure will drive the ATMwH and just send transaction data not to your bank, but to a network switchN (PLUS/VISA, CIRRUS/MASTERCARD for instance). That network switch sends it yourA your bank's switch and that switch sends it to the ibm mainframe.o  L Consider the case of your bank sending a packet authorizing the transaction.L But just after the packet was sent out on the X.25 line to the internationalL switch, your bank's switch goes down. The card network doesn't know about itJ and continues to forward the packet to the australian bank. The australianJ bank gets the authorisation and instructs the ATM to give the money to the
 customer.   K The problem is that the confirmation cannot make it way back to your bank'soN mainframe. You bank's mainframe assumes that the ATM went down and cancels theK transaction (even though the money was handed over). (There are timeouts ine? the agreements after which one assumes the transaction failed).    ------------------------------  % Date: Sun, 27 Jan 2002 22:59:38 +0100o, From: Didier Morandi <Didier.Morandi@gmx.ch>' Subject: Cross Compiler to SUN/Solaris? % Message-ID: <3C54784B.3E4B92C@gmx.ch>o  H I have a Customer who would like to know if there is a cross compiler toM translate VMS Fortran (77 and 90), Cobol, Basic and Pascal sources to compiler under SUN/SOLARIS.   Thanks,    D. --  H   ----------------------------------------------------------------------E MORANDI Consulting.  WEB: http://Didier.Morandi.Free.fr/index_us.htmlrE Pflanzschulstrasse 53, 8004 Zurich, Switzerland. GSM: +41 79 705 4670e/ 19, chemin de la Butte, 31400 Toulouse, France.e  H Disaster Recovery Plans, Computer Security Audits, DEC OpenVMS ExpertiseH On parle franais Man spricht Deutsch se habla Castellano English spoken   ------------------------------  % Date: Sun, 27 Jan 2002 17:36:18 -0600eC From: "Craig A. Berry" <craig.berry@nospam.signaltreesolutions.com>n+ Subject: Re: Cross Compiler to SUN/Solaris?o= Message-ID: <3C548EF2.9070006@nospam.signaltreesolutions.com>p   Didier Morandi wrote:n  J > I have a Customer who would like to know if there is a cross compiler toO > translate VMS Fortran (77 and 90), Cobol, Basic and Pascal sources to compile  > under SUN/SOLARIS.    E The Forte FORTRAN compiler claims to have VAX FORTRAN 5.0 extensions:-  / http://www.sun.com/forte/fortran/datasheet.html   F IIRC, Micro-focus COBOL supports just about everybody's extensions andI runs on just about everything, but I haven't looked at it in a long time.w  G But as you no doubt know, where things get interesting is when the code1F relies heavily on native RTLs. I think you'd have a porting project onG your hand regardless, and it's impossible to say how big of one withoutl" knowing what the code lookes like.   ------------------------------    Date: 27 Jan 2002 17:46:27 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)E+ Subject: Re: Cross Compiler to SUN/Solaris?e3 Message-ID: <pz4ksZ2oZ7kW@eisner.encompasserve.org>e  T In article <3C54784B.3E4B92C@gmx.ch>, Didier Morandi <Didier.Morandi@gmx.ch> writes:J > I have a Customer who would like to know if there is a cross compiler toO > translate VMS Fortran (77 and 90), Cobol, Basic and Pascal sources to compile  > under SUN/SOLARIS.  / Those compilers all support language standards.t  D If the programs were written using the standards and no VMS-specific" extensions, it should "just work".  ! Otherwise, real work is required.k   ------------------------------  % Date: Mon, 28 Jan 2002 05:59:16 +0100 , From: Didier Morandi <Didier.Morandi@gmx.ch>+ Subject: Re: Cross Compiler to SUN/Solaris?I& Message-ID: <3C54DAA4.A199BB38@gmx.ch>  J So, if I understand you well, and it seems easy, my new question would be:  L Solaris equivalences to the RTL routines? (and SS calls, and SMG and and...)   D.   "Craig A. Berry" wrote:s > G > The Forte FORTRAN compiler claims to have VAX FORTRAN 5.0 extensions:e > 1 > http://www.sun.com/forte/fortran/datasheet.html- > H > IIRC, Micro-focus COBOL supports just about everybody's extensions andK > runs on just about everything, but I haven't looked at it in a long time.1 > I > But as you no doubt know, where things get interesting is when the codeoH > relies heavily on native RTLs. I think you'd have a porting project onI > your hand regardless, and it's impossible to say how big of one without $ > knowing what the code lookes like.   Larry Kilgallen wrote: > 1 > Those compilers all support language standards.  > F > If the programs were written using the standards and no VMS-specific$ > extensions, it should "just work". > # > Otherwise, real work is required.h   ------------------------------  # Date: Mon, 28 Jan 2002 02:44:38 GMTs1 From: "David J. Dachtera" <djesys.nospam@fsi.net> D Subject: Re: Gartner Group lies about VMS ... loses all credibility!' Message-ID: <3C54BC3A.75FCC1C8@fsi.net>    "David L. Cathey" wrote: > b > "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message news:<3C523428.D080A22D@fsi.net>... > > "David L. Cathey" wrote: > > > [snip]* > > > Then how much can you afford to pay? > >h9 > > This is exactly the point I can't seem to drive home.p > >n? > > What *I* (dba DJE Systems) can afford to pay is immaterial.  > D > But you are using that as the basis to demonstrate "unaffordable".  % No, I'm not. Read the next statement.n  J > > The issue at hand is what the *END USER* (read: my (or your) customer)I > > can afford or cost-justify, and VMS ain't it 95% or more of the time.t   Get it *NOW*???!!!  L > > Could I scrape up a grand for the CSA? Sure. It'd be a lean month or so,7 > > but I could without harming the household too much.  > >hJ > > But what would it matter? I could devote the rest of my useful life toL > > porting the kitchen sink to VMS - but if no one (or very few) can afford > > VMS, what's the point? > J > Again, what is your basis for unaffordable?  Real life or your financial
 > situation? n  G Neither - the prospects' budget, etc. God, I would give my eye-teeth to ; be able to pound that through *ANY*one's skull just *ONCE*!   @ > You seem to think your boss would be bery interested in a $10KK > system - I've shown you one.  Now are you going to make the sale, someoneh) > else, or are you going to let is slide?   D That rather depends on Compaq/OpenVMS. How many *MORE* obstacles areE they going put in the way? (Pricing? Further damaging their virtuallyrH non-existant credibility? Kill the IPF port in favor of AMD Hammer? Your guess is as good as theirs!)  " > > What can your customer afford? > >l > > Ask *THEM* !!! > D > Actually, I do ask them.  One good customer of mine would NOT haveA > anything else run his business.  He leases time on a system for J > $2000/month, and would not trade it for a Windows solution for anything.K > Even with me consulting, he's very happy with his TCO (which is on top of = > the lease price), since very rarely does anything go wrong.t > O > You don't look at an OpenVMS sale as a 1-time cost.  If you do, and if that'siL > the only perception you give your customer, OpenVMS will lose and you will > lose.c  ? The one-time effort is getting the bean-counters and purchasingmH authorities past the initial "sticker shock" and demonstrating the lowerB TCO over time. Get past that, and the rest is a virtual cake-walk!  L > OpenVMS wins when you can demonstrate to the customer that the TCO for the( > system far outweighs the 1-time cost.   " Exactly my point. Tried it lately?  " > The reliability, stability, easeJ > of management (i.e. it doesn't take 2 MSCE's just to keep the box up) isJ > where OpenVMS wins.  It wins beause you turn it on and forget it.  There > is no blue screen. > O > Folks in this newsgroup keep touting how reliable OpenVMS is, how trustworthy J > the O/S is, how easy it is to work with, maintain, manage, etc.  Why did > your wife want an Explorer?   E Dangerous territory there, cowboy! ...and I'm *NOT* going to say thats again!   > [snip]> > Don't ask if your customer can afford the price of the box.   7 Wow! You REALLY haven't tried that lately, have you??!!    > As your customerP > how much his business depends upon the system being available to his employeesH > and customers.  Ask him how much a reboot during prime time costs him.  F Experience from the field shows it's a price they're willing to pay asE compared to getting over that acquisition hump. "Unexpected" expensesoF are taken as a fact of life with NT/W2K these days. No one understandsF what a stable, reliable system is anymore - it's been to long. Most ofH them didn't "grow up" professionally in the same era as stable, reliable systems.  G As I said last night - that's all the effort I'm putting into this. I'mm done after this.  > ...and before you go complaining any more about how much *OUR*C complaints are damaging OpenVMS, give some thought to the degree of F damage being done to OpenVMS by all this anti-"affordable" rhetoric. IF envision some folks going to their management saying, "See, boss? EvenD the folks who brought OpenVMS hobby into existence and facilitate itG don't support affordable commercial licensing for OpenVMS. Guess CompaqeE just doesn't want our business. So, we can scratch OpenVMS, Alpha andtB probably Compaq as well off of our list. I think Dell's prices are0 better, anyway, even if support isn't as solid."   Cuts both ways, doesn't it?a  C They may even go on to say, "Y'know, this Dachtera asshole actually D thinks he can do without M$ all together, even though he posts usingF Netscape on W/95. Maybe we should call HIM and see if he's all talk or. if he can put his systems where his mouth is!"  H DJD here, over and out. I'm gonna go see if I can crank this puppy up onD Mandrake Linux and run Netscape there, instead. Koffice looks prettyH good these days! Maybe practicing what I preach will start bringing some new business my way.   --   David J. Dachterab dba DJE Systemsn http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/a   ------------------------------   Date: 28 Jan 2002 01:26:00 GMT( From: peter@taronga.com (Peter da Silva): Subject: Re: historical evidence of what went wrong at DEC2 Message-ID: <a329b8$1v8t$1@citadel.in.taronga.com>  $ In article <vvib2a.311.ln@luciferF>,- Mike Meredith at home <hmv@port.ac.uk> wrote:dP >> I *have* had a lot of people complaining that they can't access the web pagesP >> they just uploaded with an "htm" extension because they typed "html" in theirL >> browser. Or vice-versa. Plus "tif" versus "tiff" and "jpg" versus "jpeg".  
 >See above :)a  I Just had another one. User sent me some web pages in a zip file to put up J on one of our servers. Given this recent discussion, I asked him if he hadL everything in the right case, he said "yes, Frontpage took care of that, and the home page is index.html".7  J Apparently Frontpage DOES take care of that, but I had to go add a symlink: on my Apache server because the home page was "index.htm".   *sigh*  I I think I'll look up this "apache spell check module" you referred to, if  you weren't joking about it.   -- i@ Rev. Peter da Silva, ULC.	                                 WWFD?  F "Be conservative in what you generate, and liberal in what you accept" 	-- Matthew 10:16 (l.trans)n   ------------------------------  # Date: Sun, 27 Jan 2002 20:10:14 GMT " From: "Hans Vlems" <hvlems@iae.nl>$ Subject: HP 35470 DAT drive question0 Message-ID: <G8Z48.1521$A3.7684@typhoon.bart.nl>  K HP documentation explains that the 35470 drive may be made to work on  "DEC  3100F and DEC 5000 workstations, running Ultrix and VMS". I tried the switch settings mentioned, G and several combinations. None worked. Initializing a blank tape always  results in a verym> long wait and the command stops with a "parity error" message.K There are two LED's on the drive, both green (briefly amber when the systemh is switched on).J When the INIT command is issued the right LED flashes approx. four times a second.cK I run VAX/VMS 7.2, no patches on a VAXstation 3100/GPX, hardware model typeo 59.   / So is this a Unix only device, or is it broken?e  
 Hans Vlems       BTW SCSI_INFO returns: $! $! SCSI_INFO V1.1n $!K $!      Copyright (c) 1995 by DIGITAL Equipment Corporation, Maynard, Mass. L $!      All Rights Reserved. Unpublished rights reserved under the copyright" $!      laws of the United States. $! $! Issuing INQUIRY QIOW  $!# $!   SCSI Inquiry Data (40 bytes) :e $!= $!      000:  01 80 02 02 23 00 00 18 48 50 20 20 20 20 20 20I= $!      010:  48 50 33 35 34 37 30 41 20 20 20 20 20 20 20 20 % $!      020:  31 31 30 39 33 32 34 30a $!2 $!      Peripheral Device Type : Sequential Access $!% $!      Peripheral Device Qualifier :- $!F $!              Peripheral device type supported & connected to target $!" $!      Vendor Identification : HP $!( $!      Product Identification: HP35470A $!$ $!      Product Revision Level: 1109 $! $!      ANSI Information:@ $!" $!              Complies to SCSI-2 $! $!      ISO Information: $!D $!              0 - does not claim ISO SCSI (ISO IS 9316) compliance $!   ------------------------------  # Date: Mon, 28 Jan 2002 03:49:46 GMTp9 From: "gary beeching" <garybeeching<rem>@Cableinet.co.uk> * Subject: Re: I/O Performance on Alpha 41009 Message-ID: <uT358.17042$wF1.2137737@news1.cableinet.net>s  H Brian, try out perfect cache from http://www.raxco.co.uk/fs_openvms.html  K and no - i dont work for them, but I have experienced 100-500% gain in readt performance.   ------------------------------  # Date: Mon, 28 Jan 2002 01:01:44 GMTe) From: rob.buxton@wcc.govt.nz (Rob Buxton)o& Subject: Re: mount problems on cluster2 Message-ID: <3c54a107.2425912647@news.wcc.govt.nz>  6 I'm having a similar problem, but it happens randomly.D I'm only mounting a single disk in sylogicals and that disk contains- the UAF Files and command command procedures.c  E I have a mixed Cluster, two VAX4000-700s and two Alpha DEC 3000s. The D VAX4000s have a DSSI bus that connects two HSDs. It is on these that  the UAF and common files reside.  @ Every now and then when booting one of the Alphas the shadow set7 containing the UAF and common files will fail to mount.   B My sylogicals.com file loops round and repeatedly does a sysman ioD autoconfigure and attempts the mount again but to no avail. Gives up after 5 attempts.e  A I've just added more diagnostics to the com file to try and checke' where it's going and why it's failing. t  A If I halt and reboot the machine it usually works fine the second. time. B I know I'm accessing the HSD disks because the Quorum disk is also3 served from the HSDs and the Alpha finds that okay.   E Why it works sometimes and not others is the biggest mystery. I'm notg used to VMS being random!o    4 On Sun, 27 Jan 2002 01:55:02 +0000, Robert DiRosario! <rdirosario@starpower.net> wrote:   2 >I'm having problems mounting disks on my cluster. >c- >If I reboot a node and then do "show dev /m"p4 >some of my DSSI disks show up with "(remote mount)"1 >listed for the volume label.  If I try to access / >the disk I get errors.  I can access the diskso >just fine from other nodes. >r- >If I issue the mount commands on a terminal d >everything works fine.  >p3 >If I power down everything and boot all the nodes i% >from scratch, everything works fine.o >e3 >Why are the DSSI disks not being mounted correctlya >from sylogicals.com?  c >o >In sylogicals.com I have: >a, >$mount $2$dia1  disk_d1  /cluster /noassist, >$mount $2$dia2  disk_d2  /cluster /noassist, >$mount $2$dia10 disk_d10 /cluster /noassist, >$mount $2$dia11 disk_d11 /cluster /noassist, >$mount $2$dia12 disk_d12 /cluster /noassist, >$mount $2$dia13 disk_d13 /cluster /noassist >o1 >$mount m48$dka200: disk4      /cluster /noassistc1 >$mount m48$dkb200: disk1      /cluster /noassist 1 >$mount m48$dkb300: disk2      /cluster /noassist 1 >$mount m48$dkb500: disk3      /cluster /noassist  >m# >The cluster system disk is $2$dia0A >i+ >I have nodes A (4000/106A), B (4000/106A),r. >C (4000/105A) and M48 (VAX Station 3100/M48). > - >Nodes A, B, and C, disks dia0, dia1 and dia2n! >and a HSD10 are on one DSSI bus.> >t3 >Nodes A and B, disks dia10, dia11, dia12 and dia13 * >and a 2nd HSD10 are on a second DSSI bus. >i1 >Node C and disks D20, D21 and D22 are on a thirdu2 >DSSI bus, but I haven't added mount commands for  >those drives yet. >I1 >Node m48 is on ethernet and works fine, but it'sw6 >turned off for now.  It has it's page and swap files 5 >on a local SCSI disk, and when nodes A, B or C boot y6 >they run sypagswpfiles.com and try to use m48's page 3 >and swap file.  (A quick search of google turns upn5 >satellite_page.com as a solution for old versions of 5 >VMS, but I'm using 7.1.  Is the node name defined byn" >the time sypagswpfiles.com runs?) >l >Thanks  >Roberta   ------------------------------  # Date: Mon, 28 Jan 2002 01:41:57 GMT - From: "John E. Malmberg" <wb8tyw@qsl.network>o& Subject: Re: mount problems on cluster* Message-ID: <3C54B089.5080906@qsl.network>   Rob Buxton wrote:w  8 > I'm having a similar problem, but it happens randomly.F > I'm only mounting a single disk in sylogicals and that disk contains/ > the UAF Files and command command procedures.-    + Please see my earlier reply to this thread.:     G > I have a mixed Cluster, two VAX4000-700s and two Alpha DEC 3000s. TheiF > VAX4000s have a DSSI bus that connects two HSDs. It is on these that" > the UAF and common files reside. > B > Every now and then when booting one of the Alphas the shadow set9 > containing the UAF and common files will fail to mount.  > D > My sylogicals.com file loops round and repeatedly does a sysman ioF > autoconfigure and attempts the mount again but to no avail. Gives up > after 5 attempts.     C Putting a SYSMAN IO AUTOCONFIGURE in sylogicals is unneeded, as it 99 normally is run by the operating system as it is booting.R  H Having it there may actually make your problem worse, as it is spending H more time probing the various busses on your system, which disrupts the $ I/O going on in the boot of OpenVMS.    0G > Why it works sometimes and not others is the biggest mystery. I'm nots > used to VMS being random!C    G The randomness is in the various timings involved in getting the disks e powered up and on line.r  I Many disk subsystems will not spin all disks up at once.  This may be to c> prevent the surge current from overloading the power supplies.  C The base operating system of OpenVMS does not normally require the  ; additional disks to start up, so it does not wait for them.   4 In my previous response to this thread, I mentioned E sys$examples:mscpmount.com as a starting point for building your own u mount procedures.e     -Johnt   wb8tyw@qsl.network Personal Opinion Onlyu   ------------------------------  # Date: Sun, 27 Jan 2002 23:31:02 GMTe- From: "John E. Malmberg" <wb8tyw@qsl.network> < Subject: Re: VMS system on the sourceforge.net Compile Farm?* Message-ID: <3C5491D9.3080506@qsl.network>  * Alan Winston - SSRL Admin Cmptg Mgr wrote:  ) > In article <3C53870D.53EF8F2A@ACM.org>,60 >"C.W.Holeman II" <cwhii@ACM.organization> wrote > C >>Has anyone looked into having a VMS system on the sourceforge.net@ >>Compile Farm?r >>F >>	https://sourceforge.net/docman/display_doc.php?docid=762&group_id=1 >> > E > I looked up that page, and I see a couple of problems right away.  cK > The way the compile farm is set up you need SSH to log in and SCP to get lM > your code there and retrieve your binaries.  (User-id and password are alsoaN > fhe same as on other sourceforge projects, so although they don't explicitlyO > say it, they probably need YP/NIS or some other single-sign-in authentication O > xsystem, which may not run on VMS.)  Compaq has no SSH or SCP and no plans to O > provide one; Multinet/TCPware have SSH and (I believe) SCP coming out in 4.4.sM > Although no money changes hands in the sourceforge project, it doesn't seem  > like legit hobbyist use.    E I do not know the specifics of SourceForge, but in general the build dD farm is used to make sure that when someone makes a change, that it  builds on all platforms.  A It gives some feedback, to the other developers, but if only one rH platform is broken by their change, and they do not get feedback from a 8 maintainer for that platform, the build may stay broken.  F This also means that someone must be actively maintaining the OpenVMS C port of the applications being built.  If the application will not  I currently build on OpenVMS, none of the other developers are going to be s concerned if it does not build.b  C Usually the machines on the build farm are owned by individuals or lI companies.  They automatically upload files from the CVS or other change fE control library, and then copy back the build log and test log files.x  H The uploads from the build farm to what ever central server coordinates @ them consists usually of the log file from the build and of any & validation testing of the executables.  I Normally source code changes go through the normal change control system.r  I The system doing the build will get it's source by monitoring the Change tE Control system like CVS for changes, or other methods like the RSYNC n program.  I A secure channel is used from the central server to the build machine as hE a trigger to synchronize it's sources to a specific revision, do the   build and upload the logs.  D I would assume that the security is needed on the central server to  prevent pranks and accidents.a  B On the build machine, you need special security to make sure that I someone has not checked in malicious code into the source.  Particularly rE a problem if you need to test a network server program that must run o with privileges.  F So right now, anyone with a valid hobby or commercial OpenVMS license 6 can set up to download the code and do nightly builds.  I Once those people can demonstrate that they can keep the OpenVMS variant  F of the application up to date, if SSH and SCP are not available, they F may be able to make other arrangements with the build farm maintainer.I Until then, they can submit build breakers to the maintainer of the code i* through the same methods as everyone else.  F I do not know if these capabilities can be added to OpenVMS by use of 0 the OpenSSL package.  See http://www.openssl.org  C SCP seems to be available in FISH, http://www.free.lp.se/fish/ and v< according to the link, FISH is being superceded by BAMSE at  http://www.free.lp.se/bamse/    I I am not a lawyer, and have not played one on TV, but IMHO, I do not see lA how posting the logs of a build to the public violates the hobby  @ license, as long as building them is not a commercial operation.I Even providing the binaries that way is really no different than placing m1 them on any other FTP site for free distribution.y     -Johnt wb8tyw@qsl.network Personal Opinion Onlya   ------------------------------  # Date: Sun, 27 Jan 2002 20:41:16 GMT 4 From: Tim Llewellyn <tim.llewellyn@blueyonder.co.uk>. Subject: Re: When to start a new c.o.v thread?0 Message-ID: <3C5464FF.5866A7F1@blueyonder.co.uk>   Wayne Sewell wrote:c > R > In article <3C50A2AE.32B5D217@ACM.org>, "C.W.Holeman II" <cwhii@ACM.org> writes:Q > > Since the above item provoked my posting is it correct to continue on the old 
 > > thread7 > > or is it correct to do as I did and make a new one?t > >t > O > If you survey the current subject lines, you will see that many of us start amQ > new subthread by creating a new title that also contains the old title prefixedwL > with "(was: ".  This allows people to realize that it is a new thread, butP > there is still a tie to the old thread.  Sometimes this happens several times. > 9 > For instance, if you followup to a post with a title ofe > * > "Compaq marketing policy drives me nuts" > O > and it diverges off to something totally unrelated (as they often do), change  > the title as follows:  > E > "I love/hate vegemite (was: Compaq marketing policy drives me nuts"h > O > If your newsreader allows editing of subject lines, this is trivially easy tor > do.a >   R yup, Wayne, thats the way I try to do it, must admit to being lax recently though.   regards    -- ] Tim.Llewellyn@cableinet.co.uk  n  C Standard disclaimer applies. My views in no way represent those of t! my employers or service provider.t   ------------------------------    Date: 27 Jan 2002 20:18:56 -0800, From: tony.cheung@asiayeah.com (Tony Cheung)> Subject: Re: writev() crashes with access violation on OpenVMS= Message-ID: <f9dc0a5a.0201272018.216b74c2@posting.google.com>d  [ Colin Blake <colin@theblakes.com> wrote in message news:<3C5149B8.91D9CC1@theblakes.com>...h& > socket does return an fd on OpenVMS. > I > The bug is in the CRTL's writev function and only occurs if the fd is adI > socket. I'm not sure when it got fixed or if that fix is in an ECO yet.  >  > Colin.   Thanks.t  C I think I am already using the latest versions, OpenVMS 7.3 and the. Compaq C++ 6.3-020.m   I've tried,   + writev(decc$socket_fd(sockfd), &iov[0], 1);    However, it also crashes.   F It looks like I have to avoid writev() and uses write() instead on the OpenVMS for the time being.o  7 Anyone else who could give us some thoughts? Thank you.    Tony Cheungh   ------------------------------  % Date: Sun, 27 Jan 2002 13:49:27 -0500c: From: shannon@news.widomaker.com (Charles Shannon Hendrix)9 Subject: Re: Younger recruits versus experienced veterans . Message-ID: <n3i13a.gsg.ln@escape.shannon.net>  ' In article <3C4EF7E4.555783CA@ev1.net>, + Charles Richmond  <richmond@ev1.net> wrote:o  > Charles Shannon Hendrix wrote: > > ; > > In article <name99-1901022044170001@handma2.apple.com>,w+ > > Maynard Handley <name99@mac.com> wrote:u > > N > > > Apple, of course (and I am quite serious here) no longer do this either.M > > > There are still ten variants coded up, but they are all showed to StevesP > > > Jobs, not "the public" who then decides what he wants. Hence the continualL > > > fury by almost everyone but Steve Jobs at the UI decisions in MacOS X. > > I > > I can't figure out why Steve Jobs gets so much credit.  I have yet totJ > > read anything about him that doesn't show him to be a perpetual whinerH > > who had little to do with creating what Apple is or has, and in fact= > > is responsible for some of Apple's worst characteristics.v  F > Jobs drove the Mac team *very* hard to get the product out. Read the( > book _Insanely Great_, by Steven Levy.  D I have, as well as "Inside Apple" and "The Journey is the Reward".    G > > He does bring in good people, to his credit, but they've more often 7 > > worked around him or in spite of him than with him.   = > You have to be aware of his RDF...reality distortion field.-  - ..without getting sucked into it, right... :)"  K > > I'm sure there is a lot I don't know, but what I do know is frustratingn( > > because of the messes he has caused. > > @ > He is messy and tempermental...but Apple would *not* be Apple  > without his influences...   @ True, but then it might be something better.  We can never know.  E According to everything I have read, most of the great ideas were notkG his.  In fact, his ideas were often the cause of failure, increased and C unecessary costs, and things the users did not really want or need.,  G It's easy to see how he both created and killed NeXT, by setting up the F creation of a truly great system, only to then kill it by insisting onE oddities that garanteed difficult introduction, and priced it far out-/ of reach of nearly everyone who might want one.0  B I suppose he or someone like him was necessary, but I just wish heE would come back down to Earth now and then.  I think in some ways, he F is different now.  At least, it seems Apple is trying harder to do the right thing.  I hope so anyway.a   -- r  H UNIX/Perl/C/Pizza__________________________________shannon@widomaker.com   ------------------------------  % Date: Sun, 27 Jan 2002 13:58:19 -0500a: From: shannon@news.widomaker.com (Charles Shannon Hendrix)O Subject: Re: Younger recruits versus experienced veterans ( was Re:     The dems. Message-ID: <bki13a.vih.ln@escape.shannon.net>  . In article <a2r9sh$85a$1@nyheter.chalmers.se>,+  <alapaa@NOOOOSPAM.math.chalmers.se> wrote:nJ > In comp.arch Charles Shannon Hendrix <shannon@news.widomaker.com> wrote:; > > In article <name99-1901022044170001@handma2.apple.com>, + > > Maynard Handley <name99@mac.com> wrote:  > M > >> Apple, of course (and I am quite serious here) no longer do this either.iL > >> There are still ten variants coded up, but they are all showed to SteveO > >> Jobs, not "the public" who then decides what he wants. Hence the continual0K > >> fury by almost everyone but Steve Jobs at the UI decisions in MacOS X.I > I > > I can't figure out why Steve Jobs gets so much credit.  I have yet to J > > read anything about him that doesn't show him to be a perpetual whinerH > > who had little to do with creating what Apple is or has, and in fact= > > is responsible for some of Apple's worst characteristics.  > G > Well, then you should do some reading. I am not a MacOID, so I do not   ! I have, extensively.  I still am.   H > follow all the events reported in the press about what happens at/withB > Apple. What I do know is that Jobs and Wozniak basically createdD > Apple. Even more importantly (as for Jobs' competence) when he was  G Apple was created by Wozniak.  He did just about all of it, Markula did<' the rest.  Jobs was often in the way.  P  D Now, had Jobs not been the zealot he was, it's true that Apple wouldG not be here.  Wozniak had no interest in selling the ideas.  That makesfF Apple Jobs' idea, but it was almost entirely the creation of Wozniak's8 engineering and the business work of Markula and others.  3 Jobs was the pusher, the guy who did the selling.  r  D I'm not trying to say he was not necessary, just that he gets creditE for things which were not his doing, and a lot of people overlook thei damage he has done.   G > fired from Apple he created the NextStep OS and computers, which even E > by the standards of the OS:es today (more than 10 years later) must D > be considered far superior to most when it comes to the technologyH > of the future, OO. Imagine where the industry could have been today if; > we had had the pure OO philosophy of NextStep and pure OOnG > componentware sold by third-party vendors instead of WinBloat 3.X andiF > VBX components in the early nineties... <sigh> Furthermore, NextStepG > had a lot of other nice features such as true WYSIWYG on the display a> > using display PostScript, alpha channels (transparency) etc.  G Oh, it would definitely be nice.  However, you are talking about thingsnH which Jobs didn't create himself.  That's kind of my point: he's gettingF credit for stuff as if he did it all himself, which isn't true at all.  F There were teams behind him and some of them are still with him today.( It seems Jobs' RDF hides them sometimes.     -- '  H UNIX/Perl/C/Pizza__________________________________shannon@widomaker.com   ------------------------------  % Date: Mon, 28 Jan 2002 01:02:22 -0000s/ From: Michael Zarlenga <zarlenga@conan.ids.net>aB Subject: Re: [Q] Why is there the limit of 8 levels of directories/ Message-ID: <u598ou6cm3gt9f@corp.supernews.com>   2 Bill Gunshannon <bill@triangle.cs.uofs.edu> wrote:F : All this reminds me of my days SysAdmining Pr1mes.  It seems someoneE : who had never worked worked with tree style directories was curiouso@ : about how far down you could actually go.  He never found out.F : He wrote a little program that created a directory, moved to it, andF : repeated.  It was impossible to tell how deep he was when it crashedD : the system.  It was also impossible to do anything with this superC : deep tree.  You couldn't even delete it.  We ended out backing updE : everything else ont he disk and re initializing the whole thing andoF : then restoring all the user files.  In those days, on that hardware,@ : it took more than a a day and the other users were not amused.  ; : Maybe having a limit below critical mass is a good idea!!   8 Or building a system that doesn't crash simply because a user created a sub-directory   -- 2 -- Mike Zarlenga   ------------------------------   End of INFO-VAX 2002.053 ************************- >Nodes A, B, and C, disks dia0, dia1 and dia2n! >and a HSD10 are on one DSSI bus.> >t3 >Nodes A and B, disks dia10, dia11, dia12 and dia13 * >and a 2nd HSD10 are on a second DSSI bus. >i1 >Node C and disks D20, D21 and D22 are on a thirdu2 >DSSI bus, but I haven't added mount commands for  >those drives yet. >I1 >Node m48 is on ethernet and works fine, but it'sw6 >turned off for now.  J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    J    