eli@pws.bull.com (Steve Elias) (09/06/90)
In article <33571@cup.portal.com> DeadHead@cup.portal.com (Bruce M Ong) writes: >i wrote: >>why not put your image file on a hard disk? why use a floppy drive >>or slow network drive for anything the least bit time critical? >> > I was evaluating the boards, not running a system with them. > And yes, that's how I got around the problem with slow server - copy the >images to a local disk before sending them out. in my smiley opinion, you'd probably run into trouble under *any* OS if you try sending a fax from a floppy drive. floppy access on PC-bus machines is notoriously nasty and difficult to deal with from the OS, especially if the PC-bios is used. > Another weird problem with MSDOS based file system for these fax >cards (not just gamma link's) running real time, is that if you are sending >a file that is in a directory with A LOT of files, searching thru the FAT >table will take sometimes minutes, and opening of files, etc. slows down >tremendouisly. minutes?!?!? i'm crossposting this to comp.sys.ibm.pc. were you running on a 4.77 Mhz XT with a very big disk drive? is this FAT access problem something you saw on hard disks as well as floppies? the FAT access on big DOS disks does slow things down a bit, as i've seen similar problems on multichannel (4-6 channels) PC-AT DOS-based voice mail systems. but the FAT access latencies i've seen in this situation are on the order of .5 or 1 second at max -- not many seconds. so, i'm wondering about the cause of the problem you described. any ideas, yall? >That could cause trouble, too. I guess not a whole lot people >are running these fax boards in a real real serious, and industrial >environment with heavy loads during peak hours. "fax-server" applications & systems are not as abundant as voicemail apps & systems, but the technical problems are very similar. perhaps you ought to contact a faxboard & system manufacturer that has lots of experience with voice mail as well! :) >>how about fixing it yourself by using a local hard disk? >>imho, it's YOUR problem, not Gammalink's. >> > Hey. it's MessDos. <grin> reiterating: floppy access on PC-bus machines is a big NO-NO if you are doing time critical multitasking, especially multichannel realtime stuff. regardless of OS, imho. as for MessDos, i can't stand it, personally, but it is possible to do multichannel voice mail under DOS, so multichannel fax is quite possible as well. Brooktrout now has a DOS multichannel fax toolkit, although i haven't used it. >>obviously in the running. any dialogic fax users out there? > > I dont think the dialogic stuff is out yet. their brochures are out, though! :) >>why not use 386 unix instead? > > How many multi-channel fax cards for the PC have 386 unix drivers? >I'd like to compile a list for that.. there is JTFAX, i think ( you dont >need a driver). There is Brooktrout. There is err... Dialogic (when it comes >out.) Nobody is using the yamaha chipset in any of these boards I dont think. >(Which makes the modems pretty expensive). we'll see how *nice* the products built with Yamaha chips end up being. but, i guess you're right that there aren't all that many fax boards which have unix drivers. but there are more coming. some of these DOS-fax companies are finally realizing that Unix is more than a four letter word. > And there are the serial fax >modems - but the software they supply all run under DOS, and we still dont >have a standard way of communicating with these serial fax modems. ah, but we do -- it just hasn't been implemented by most serial-fax-device manufacturers. check with Datarace -- they were all over the creation of the serial-fax standard. naturally, i can't recall the standard name. something like EIA229/TRsomething. (really helpful, eh?) > Hey. But that's why we are here. Sometimes I think the confusing >computer industry creates careers for people - just as the confusing >social trends after the 50's create careers for shrinks and people who >write how-to books. But... I am not complainin' (about the confusions in >the industry creating careers for people, at least). Hey now! i agree completely. let's hear it for job security! :) -- /* eli@pws.bull.com 617 932 5598 fax 508 294 0101 */
DeadHead@cup.portal.com (Bruce M Ong) (09/07/90)
> >> Another weird problem with MSDOS based file system for these fax >>cards (not just gamma link's) running real time, is that if you are sending >>a file that is in a directory with A LOT of files, searching thru the FAT >>table will take sometimes minutes, and opening of files, etc. slows down >>tremendouisly. > >minutes?!?!? i'm crossposting this to comp.sys.ibm.pc. were you >running on a 4.77 Mhz XT with a very big disk drive? is this FAT >access problem something you saw on hard disks as well as floppies? > >the FAT access on big DOS disks does slow things down a bit, as i've >seen similar problems on multichannel (4-6 channels) PC-AT DOS-based >voice mail systems. but the FAT access latencies i've seen in this >situation are on the order of .5 or 1 second at max -- not many >seconds. so, i'm wondering about the cause of the problem you >described. any ideas, yall? > guess I should clear thing up a bit :). The environment was a 25 mhz 386, running DesqView with a process creating 7000 files in a directory 3 levels down from the root, each file 1000 bytes long. At the end when it got to about 6500 files, it got painfully slow. I think it took teh system more than 12 hours to do the task. Why would I want to do such a deranged thing? I was trying to prove to somebody who wanted to run two GammFax GFDCP processes under desqview that it would not work. I will spare the net with the rest of the gory details... > >> And there are the serial fax >>modems - but the software they supply all run under DOS, and we still dont >>have a standard way of communicating with these serial fax modems. > >ah, but we do -- it just hasn't been implemented by most >serial-fax-device manufacturers. check with Datarace -- they were all >over the creation of the serial-fax standard. naturally, i can't >recall the standard name. something like EIA229/TRsomething. >(really helpful, eh?) > Why do I have a feeling that this standard will be met by the fax modem community with resounding silence... just like the one from intel... >/* eli@pws.bull.com 617 932 5598 fax 508 294 0101 */ /bruce deadhead@cup.portal.com