Malcolm,
On 2005 May 9 , at 21.15, Malcolm J. Currie wrote:
> Just took a look at SUN/248, reminded by the recent exchange. Looks
> good. I spotted some typographical errors, and have one important
> structural suggestion.
Thanks for these.
I've put a new version at
<http://www.astro.gla.ac.uk/users/norman/star/sun248/>
> The basic instructions should be in a short separate section with
> suitable words in its title, such that search engines will readily find
> it. Something like "Generic build and install of Starlink software".
> "Portable" now I think of it might be better than "Generic". You get
> the idea. When I reached Section 2, I was wondering where the
> instrument-independent information was.
I see what you mean. I think I've fixed this, by renaming
`Introduction' to `Building and installing Starlink software: generic
instructions' and tweaking slightly. Does this serve, do you think?
> Does this system build the STARJAVA or just the Classic applications?
> The scope be stated clearly in the Introduction and Abstract.
I made a couple of changes in revision 1.2 following from this and from
Mark's observations.
> I think project meaning [Starlink] Project should begin with a capital
> letter.
Done.
> You know I prefer co-ordinate, the English spelling.
Well, yes, but I'm a little more old-fashioned, and deplore such late
19th century innovations as that intrusive hyphen. In any case, we
know it's _really_ supposed to be coördinate, but diaereses are seen as
a bit of a faff, these days.
> Section 1, para. 2, line 1. On "These are" is more "These were".
> Since we've not made any releases under the new system, can we be sure?
> They used to be biannually under the CD releases, but not since ISOs
> were made. Anyway, upon re-reading I think I'll pass the
> optimistic sentence. If we're closed down, we'll have to make changes.
Yes, I swithered about this as well, and came down on the optimistic
side, too.
> para. 2, line 2 "we target". I'd prefer "Starlink targets".
Me too, I think.
> Section 1.1. Not everyone knows what CVS is. A brief explanation and
> link would be handy.
I've added a pointer -- is this enough, do you think?
> Section 2.1, line 1/2. "Apple doesn't include g77 in its
> distribution."
How interesting -- I've always naturally thought of `Apple' as plural.
I realise I also think of `Sun', `DEC' and so on as plural. Hmm: I
realise I think of `Microsoft' as monolithically single. Companies and
organisations I like I apparently think of as a humane assemblies;
organisations I don't I think of as oppressive behemoths. Do you feel
particularly strongly about this one?
> Section 2.3, para 1, line 3 s/,././
Your optician would be proud of you.
> Section 2.6, para. 1, lines 1--2. What's this tetex 1.* package? If
> that's a version number, you need a box around it so the name and
> version are on the same line. I couldn't work out what it was on
> first reading.
Fixed.
> Item 2. s/newly-built/newly built/ because it's an adverb ending in
> "ly". I learnt this from the Chicago Manual of Style two decades ago.
Ah, but `newly-built' is a compound adjective: omitting the hyphen here
would surely lead to parsing errors.
I greatly enjoy reading style manuals, and identifying their errors
(viz, all disagreements with me, which must be either insane or
mischievous).
> Section 3. There needs to be more consistency of use of \tt for "{\tt
> mk} script", e.g. para. 2, line 1 and Item 1. I did wonder if there
> were other instances where you needed a \tt where you're referring to
> actual commands, e.g. f95 on Section 2.3, para. 3.
I've swithered about this variously here and very variously in SSN/78.
I somehow feel that when I'm referring to a particular command I should
have it in \tt, but when I'm referring to f95 in general, I shouldn't,
but I can't formulate this into any reasonable general principle.
> Item 1 missing that fullstop again. At least you're consistent!
Distressingly so.
> Section 3.2, final para. s/top level/top-level/.
Now _this_ one, I agree with.
> After reading Section 3.4 I wondered (playing the naive user), if it
> were possible to build against static libraries, and if so, how.
I've rewritten this short section. What do you think?
A bracing exercise after a rather depressing day. Much thanks, Horatio.
Norman
[you might find
<http://www.astro.gla.ac.uk/users/norman/note/2005/singular-data/>
entertaining (though a handy gin might be advisable)]
--
----------------------------------------------------------------------
Norman Gray : Physics & Astronomy, Glasgow University, UK
http://www.astro.gla.ac.uk/users/norman/ : www.starlink.ac.uk
|