wnp@dcs.UUCP (Wolf N. Paul) (06/05/88)
In article <1248@spdcc.COM> dyer@spdcc.COM (Steve Dyer) writes: >I would like to know whether Elan's troff package >for the PC runs acceptably well in a DOS environment. Acceptably >well might mean reasonably fast as well as not requiring enormous- >sized intermediate files in lieu of true pipes: the file size >expansion of a TBL --> ditroff -> output-filter --> LPR: sequence >is frightening, given any moderately-sized input file. Well, DOS implements pipes as intermediate files on disk, and since it is not multi-tasking, ALL the output of one program is written to disk before the next program in the pipe starts. The only whay ELAN could have changed that would be to provide a replacement for DOS, i.e. a different operating system. Thus, a pipe as shown in your posting would create temporary files for the output from tbl, ditroff, and the postprocessor. The best way to handle that would be to run under some shell which lets you specify the location of pipe temporaries (like the MKS Toolkit shell) and to specify a ram disk set up in a few MB of EMS memory. -- Wolf N. Paul * 3387 Sam Rayburn Run * Carrollton TX 75007 * (214) 306-9101 UUCP: ihnp4!killer!dcs!wnp ESL: 62832882 DOMAIN: wnp@dcs.UUCP TLX: 910-280-0585 EES PLANO UD
dyer@spdcc.COM (Steve Dyer) (06/06/88)
In article <107@dcs.UUCP>, wnp@dcs.UUCP (Wolf N. Paul) writes: > Well, DOS implements pipes as intermediate files on disk, and since it > is not multi-tasking, ALL the output of one program is written to disk > before the next program in the pipe starts. > > The only whay ELAN could have changed that would be to provide a replacement > for DOS, i.e. a different operating system. Thanks, I know about DOS's limitations. It is feasible for a vendor to build a multi-tasking driver program for the DWB suite and implement the pipe chains within that. Not quite a "different operating system", although not trivial (but not impossible either.) My question was whether they (or anyone else) has done this. -- Steve Dyer dyer@harvard.harvard.edu dyer@spdcc.COM aka {ihnp4,harvard,husc6,linus,ima,bbn,m2c}!spdcc!dyer