Talk:CYCLADES

Merge
I have merged this article with [CYCLADES Network]. Three References to published papers added. Rdmoore6 01:01, 5 September 2006 (UTC)

Proper capitalization
The original CYCLADES documentation always capitalized both CYCLADES and CIGALE. We should make sure to always do the same. Noel (talk) 05:24, 12 August 2007 (UTC)

Article mistake
A somewhat incorrect point which I removed from the article:


 * The successful development of this architecture in the CYCLADES network influenced the separation of the TCP and IP layer in the Arpanet architecture.

No; CYCLADES influenced the move of reliability (checksum/acknowledgement/ timeout/retransmission) into the hosts. The split of IP and TCP, to enable the addition of UDP, came about from the need to support media applications which didn't need reliability, i.e. packet voice, although the fact that the network didn't do reliability did make this split possible. Noel (talk) 05:24, 12 August 2007 (UTC)

PS: It was the Internet that was influenced by CYCLADES, not the ARPANET. The latter predated CYCLADES (and so could not have been influence by it). Noel (talk) 04:38, 27 May 2009 (UTC)

Influence
Whether the ARPANET or CYCLADES was the single biggest technical influence on the Internet is a matter of opinion which is hard to objectively settle. Part of the problem is that the ARPANET predated CYCLADES, so in some sense CYCLADES is itself an ARPANET legacy. To say which had the biggest influence, I would think one has to look at the specific novel technical features of CYCLADES and compare their importance with the contributions of the ARPANET. Clearly, the novel ideas(s) in CYCLADES were moving the reliability into the hosts. Whether that matchess the contributions from the ARPANET, which was the first actual packet network, is a tricky judgement call (one I don't have an immediate opinion on, I would have to consider the matter for some time), although the notion that CYCLADES was more important than the first actual, working, packet data network would seem to me to be difficult to sustain. (For example, for almost two decades, all the major applications were pure ARPANET application clones.)

In addition, in listing the important predecessors of the Internet, one ought to include the PARC Universal Protocol work, although the extent of the contribution there is hard to measure since i) PUP was roughly contemporaneous, and ii) many ideas that appeared in both PUP and TCP/IP (e.g. separation of the internet layer from the transport) were not copied from the other, but had an adoption driven by local influences. I have left CYCLADES as 'one of the two most important predecessors of the Internet', but whether it should be 'three', or what, is probably open to debate. Noel (talk) 03:34, 25 December 2009 (UTC)


 * It's not open to debate. Wikipedia isn't editorial; it's a mass of sourced statements. The form is "in the words of (source), CYCLADES was QUOTE one of the two most important UNQUOTE" or alternatively "CYCLADES is often regarded as one of the most important etc" (source) (source). Otherwise it's just you talking, and who are you? Even if you were Louis Pouzin, you still have to source this. -Ashley Pomeroy (talk) 20:25, 24 November 2014 (UTC)

Copy of software?
Did the actual CIGALE software survive? It should be possible to run it on emulated Mitra-15 (technical manual is available from Bitsavers project). shattered (talk) 19:01, 17 March 2017 (UTC)