ViewVC logotype

Contents of /bcast-fraction-threshold/comments

Parent Directory Parent Directory | Revision Log Revision Log | View Revision Graph Revision Graph

Revision 1.4 - (show annotations)
Thu Mar 22 21:04:25 2007 UTC (12 years ago) by riso
Branch: MAIN
CVS Tags: arXiv, HEAD
Branch point for: sirocco2007-cameraready
Changes since 1.3: +6 -5 lines
Ignorovane requesty

2 %Riso: Navrhujem takyto postup: najprv spravit arXivovu verziu
3 %(odkomentoval som appendix a zakomentoval citaciu), placnut na arXiv
4 %(mozem spravit, ked to dokoncime), potom doplnit citaciu, zakomentovat
5 %appendix a submitnut ako camera ready. Je to ok? S copyrightom by to
6 %myslim malo byt koser (je to sice tak na hrane, ale ked to presadzem bez
7 %lncs stylu a bude to s appendixom a je to pred tym, ako im posleme
8 %copyright form, tak sa to asi da...)
9 - Post na Arxiv:
10 provided in the Appendix, cannot be checked in available time. I suggest
11 that those details be made
13 - co tym myslel?
14 %% netusim....
15 %Riso: asy to, ze nie a hypercube, ale the hypercube. A potom by sme asi
16 %nemali pisat d-dimensional hypercubes, ale iba the d-dimensional
17 %hypercube? No neviem, je to tak?
18 "The hypercube Q_n", there is only one, with N=2^n vertices.
20 - Checknut anglictinu:
21 %% este skusim, ale..... well.....
22 %Riso: Hm, to je asi momentalne najvacsi problem. Ale co uz, nie sme
23 %jedini, co nie su native speakri...
24 The paper is generally well written, although the English is
25 somewhat awkward at places. The limited space did not allow the
26 ---
27 The paper is marred by linguistic problems: missing
28 commas and articles; sentences too long and convoluted
29 (e.g., pg:4 l:-10 sentence "Then we show..."), singulars/plurals (e.g.,
30 pg:1 l:-11 "approach" -> "approaches"). There are too many to list. I
31 reccomend
32 the authors have the paper proofread professionally.
36 ==================== Navrhujem ignorovat ========================
38 %% tomuto nerozumiem - co tym chcel povedat?
39 %Riso: ani ja, mne sa to zda vcelku ok... navrhujem ignorovat
40 - Restructure the two last paragraphs of section 6 -- delete the empty first
41 sentence.
44 % toto je podla mna obsolete tym, ze sme presunuli dokaz z appendixu
45 p.9/10 maybe a better explanation of the purpose of set P would help
47 - mame konzistentne male n:
48 Make sure that the number of vertices N is capitalized.
50 - je to tam napisane podla mna jasne:
51 well written. Though I do not like that the results presented in
52 the abstract are in the form O(D log n) suggesting that broadcasting
53 is feasible in this time in any graph with the diameter D, while in
54 fact the studies refer only to cliques and hypercubes.
56 - zda sa mi jasne... aj tak je to iba vo footnote.
57 (4) The notation $O(1/n)$ is meaningless. Say in words
58 what you mean.
60 - Citujeme dolny odhad pre fraction threshold, tak nejake mame. Netrivialne
61 nevieme.
62 I also find hard to accept that there is no attempt to prove any
63 lower bounds to show that O(log n)-time broadcast is optimal.
65 ========================== HOTOVO ===============================
66 - natiahnut na 15 stran (nieco z appendixu naspat)
67 Navrhujem:
68 %%DONE: Theorem 1 - cca +1 strana
69 Theorem 4, zmergovat s tym intuitivnym popisom, cca +1 strana
70 %%iba som preplacol proofsk->proof
71 este nieco - neviem sa rozhodnut...
72 mame 14 stran, takze uz moc nemozme vymyslat.... skusil som presunut dokaz lemy 6...
74 (5) In the proof of Lemma 6, the reference to the proof structure of a
75 lemma in another paper is useless for the reader.
76 %% presunul sa dokaz.....
78 p.1. l.3 "high probability of failure of some component" - you
79 probably meant that there is a high probability that some components
80 may fail, the way it is written it might be interpreted as "some
81 components fail with high probability". Rewrite/disambigue.
83 %% nieco som pripisal, nie som si 100% sure, ze je to dobre....
84 - Define carefully the "arcs" of an "edge" ("opposite arc"?)
86 - Define "many" and "few" on p. 5
87 Navrhujem zredukovat a skratit popis na uplne intuitivny a dat cely dokaz
88 Theorem 1
90 - Pripisat, ze nie su ack:
91 guess that each node when sending a message must use all edges
92 incident to it. I.e., nodes cannot choose for sending only selected
93 edges. If edges selection mechanism was available one could provide
96 There is also some confusion concering n. In cliques n stands for
97 the number of nodes and in hypercubes the number of nodes is 2^n.
99 Delete "for a long time" on p.2
100 Fix "zerowith" on p.3
101 Add "r" in "fractional"
102 p.3. line 5 "zerowith"
103 p.3. above definitions: you use both N and n, why?
105 (1) There is a hidden assumption: the nodes need not only
106 to be synchronous but also to have a "coordinated start-up".
107 In fact, otherwise the two time steps (of which is composed
108 a simple round) can be off by one time unit for different nodes.
110 (2) There is confusion thoughout the paper on the meaning
111 of the symbol $n$: in the complete graph means the number
112 of nodes (i.e., $n=N$) while in the hypercube is the
113 number of dimensions (i.e., $n = \log N$). Please be
114 consistent, eg. just use $N$ to indicate the number of nodes.
116 (3) The definition of the important parameter $X$ is hidden
117 in a sentence at the end of section 2. Either emphasize it (e.g., center
118 it) or repeate it in Theorems 1 and 2.
120 p.3. maybe a table of your results would be appropriate

CVS Admin">CVS Admin
ViewVC Help
Powered by ViewVC 1.1.26