Anne & Lynn Wheeler <[EMAIL PROTECTED]> writes: > the end of the tag field. in any case, after some years, I did a > VMSG-clone implemented in REXX and XEDIT. Later, I added SMTP/RFC822 > input/output processing.
for even more drift ... early on I was doing all this email stuff. I had a 15,000 entry nickname file that eventually grew to over 25,000. I was doing a lot of early online telephone book stuff (which was also borrowed by the PROFs group). Jim Gray and I (and others) were sitting around drinking one friday night (before he left for tandem) ... back in the days of system/r (original relational/sql) http://www.garlic.com/~lynn/subtopic.html#systemr talking about what silver-bullet application could we come up with that would tempt a lot of non-computer literate employees to start using online computing (besides the quickly emerging email application) ... and came up with the idea of a highly efficient online corporate telephone book. so the ground rules were two fold ... it had to be blazingly fast, the application had to take less than one-week to implement and the administration processing gorp for keeping all the data up-to-date and distributing it ... had to take less than one-person day per month. so the online phone book routine used radix search on sorted cms files ... which gave a noticeable improvement over binary search (we actually calcualted letter frequency distribution for names and used that for a modified radix search). so what does this have to do with implementing a vmsg-clone in rexx & xedit? well default vmsg was to do sequential scan of nickname files ... and this was starting to be somewhat more time-consuming when you had a 25,000 entry nickname file (which was much smallter than the aggregate number of entries in the collected online corporate telephone directories). so a trivial addition to vmsg-clone implemented in rexx ... was to do a modified radix-search ... similar to the implementation done for the corporate telephone directories. misc. past posts mentioning the online telephone directories and/or modified radix search (using calculate letter frequency): http://www.garlic.com/~lynn/94.html#31 High Speed Data Transport (HSDT) http://www.garlic.com/~lynn/2001c.html#88 Unix hard links http://www.garlic.com/~lynn/2002e.html#33 Mainframers: Take back the light (spotlight, that is) http://www.garlic.com/~lynn/2004c.html#0 A POX on you, Dennis Ritchie!!! http://www.garlic.com/~lynn/2004p.html#13 Mainframe Virus ???? http://www.garlic.com/~lynn/2005c.html#38 [Lit.] Buffer overruns http://www.garlic.com/~lynn/2005c.html#43 History of performance counters http://www.garlic.com/~lynn/2005s.html#3 Flat Query http://www.garlic.com/~lynn/2005s.html#4 Flat Query this helped contribute to a different problem. at the time, there was supposedly an issue with the availability of 3270 terminals ... and there was requirement for division vp sign-off for internal 3270 requests and associated delays could be six months. there was this point in time when middle-management discovered that higher level executives were starting to use online email ... and all of a suddened, every middle manager in the company seemed to demand an immediate 3270 terminal on their desk. this, in turn, pre-empted the whole six month 3270 terminal provisioning process for those programmers that actually needed 3270 for things like development (effectively six months of internal 3270 terminal deliveries were pre-empted). to help breakup that log-jam, we put together a business analysis that 3-year fully depreciated 3270 capital costs turned out to be less than monthly phone expense ... and nobody was suggesting that it should require division vp approval whether an employee got a phone on their desk or not. -- Anne & Lynn Wheeler | http://www.garlic.com/~lynn/
