wls@astrovax.UUCP (William L. Sebok) (12/13/84)
> > Personally I think RTS and CTS flow control is less disgusting than XON and > > XOFF (I think XON, XOFF is disgusting). I think that flow control should have > > been done out-of-band whenever possible. > > XON and XOFF *are* out-of-band. They are control characters, reserved for > such signalling purposes, not data characters. Re-read the ASCII standards > if you don't believe me. I am quite aware of that. I realized after I had posted that followup that it would provoke a response from some pedant. The article to which I was replying did not mention ASCII. What I was objecting to was the statement that use of RTS and CTS for flow control is disgusting. I don't have the standards before me but I seem to remember that was their intended use. Whether it is or is not the intended use of these lines. I still prefer methods of flow control be used whenever possible on 8 bit lines which allow the use of the full 8 bit path. The other nice thing about use of such lines is that flow control could have been done entirely by hardware made transparent to software. This is what how I think the standards really should have evolved. -- Bill Sebok Princeton University, Astrophysics {allegra,akgua,burl,cbosgd,decvax,ihnp4,noao,princeton,vax135}!astrovax!wls
lauren@RAND-UNIX.arpa (08/05/86)
I see I should have been more explicit--my use of the term "hardware" flow control may have been misleading. What I should have said is that ethernet and X.25 environments have RELIABLE end-to-end (computer-to-computer) flow control--usually kernel based. This makes a variety of efficiencies possible. This is in contrast to, for example, standard serial line communications, where hardware flow control often represents the only reliable end-to-end flow control mechanism, particularly at higher speeds. --Lauren--