james@contex.UUCP (James McQueston) (01/12/91)
In article <9101110004.aa25179@WOLF.BRL.MIL>, butler@BRL.MIL ("Lee A. Butler") writes: > ... > was appalled (sp?) at the namespace collisions with libgl. Why couldn't SGI > have put some standard prefix like "gl_" on the front of all their routines? > ... I naively thought that SGI was going to address this issue way back when the gl's select(3) functiion collided with the unix select(2) function. Instead, they just changed gl's select() to gselect(), sidestepping the larger issue. > ... > Changing to this shouldn't cause much impact, as there could be an OPTIONAL > include file that performed name mapping between the gl_function names and > the old style function names. > ... I am compelled to second this motion whole-heartedly. Name space collisions are one of the most difficult problems to overcome when one attempts software component re-use on almost any scale. We have been bitten several times. Please, SGI, heed Mr. Butler's advice. You too, IBM. > Lee A. Butler > Aberdeen Proving Ground, MD 21005-5066 > Ballistic Research Laboratory Phone: (301) 278-9200 > SLCBR-VL-V Internet: butler@brl.mil James H. McQueston (outgoing mailer lies: reply to contex!james@uunet.uu.net)
davea@quasar.wpd.sgi.com (David B.Anderson) (01/17/91)
In article <7865@minyos.xx.rmit.oz.au> mg@ (Mike Gigante) writes: >drw900@anusf.anu.edu.au ("Drew R Whitehouse") writes: > >]In article <9101110004.aa25179@WOLF.BRL.MIL>, butler@BRL.MIL ("Lee A. Butler") writes: [stuff deleted ] >]|> was appalled (sp?) at the namespace collisions with libgl. Why couldn't SGI [stuff deleted ] In 3.3 there is a tool, collide(1), which will find namespace collisions for you (it won't *solve the problem* though). Check the man page. [ David B. Anderson Silicon Graphics (415)335-1548 davea@sgi.com ] [``What can go wrong?'' --Calvin to Hobbes]