Re: Bug#681215: Package description too detailed
Martin Eberhard Schauer wrote:
> I'm sorry to say that IMHO the package description (1) has a number of
> problems:
>
> - The description is copied from upstream (2).
Not intrinsically bad, but it explains the rest.
> - Upstream had a list of 6 items that have become one big fat paragraph
> in the description.
> - It is too detailed and it is marketing speech.
Definitely true.
>
> Description: C library for the Swiss Ephemeris
>
> Not being a physicist or astronomer one has to look up what an ephemeris
> is (3).
For a library package synopsis there's a presumption that "if you
don't know what it means you don't need to know", so I'd say this is
forgivable, but the long description should explain it.
>
> libswe0 allows programs to access the Swiss Ephemeris.
> The SWISS EPHEMERIS is the high precision ephemeris developed
> by Astrodienst, largely based upon the DE406 ephemeris from NASA's JPL.
> The Swiss Ephemeris is based upon the latest planetary and lunar
> ephemeris, DE405/406, developed by NASA's Jet Propulsion
First it says the Swiss Ephemeris is *largely* based on DE406... then
it says it's based on DE405 and 406. Can we assume the hedging in the
first is covered by the inclusion of DE405 in the second? And is one
of them planetary and the other lunar, or what? (Looks it up) No, 406
is a smaller, lower accuracy version of 405, and the Lunar Ephemeris
is technically LE405/406.
> Laboratory. The original integration, DE405, covered the years 3000 BC
> to 3000 AD and required 550 Mb of disk space. DE406 is a compressed
^^
Megabits? Upstream typo.
> version of DE405 which requires 200 MB while maintaining a precision
> of better than 1 m for the moon and 25 m for the planets. These data
^^^^^^^^^^
"Data" has been idiomatically singular (a non-count noun) since before
I was born.
> have been further compressed with sophisticated compression techniques
^^^^^^^^^^^
Repetitive bragging.
> developed by Astrodienst. The ephemeris now requires for the complete
> 6000 years only 5 Mb for all planets except the Moon, and 13 Mb for
^^ ^^
The Moon is a planet now? Pluto must be so jealous.
> the Moon. This compressed ephemeris reproduces the JPL data with 0.001
> arcseconds precision.
>
> Astrodiest's have extended the timespan of the
^ ^
What is this, astrologer's apostrophe? Oh, and would you believe
they've misspelled Astrodienst?
> JPL ephemeris by numerical integration, so that Swiss Ephemeris covers
> the years 5400 BC to 5400 AD, a total of 10'800 years. For this extended
^
An LC_NUMERIC error, and I hope they're not counting "0 AD".
> timespan the ephemeris requires 32 Mbytes of ephemeris files.
^^^^^
I'm already so tired of the word "ephemeris".
>
> All transformation steps from the inertial timeframe of the JPL DE406
Does this sentence with its much-delayed verb maybe sound reasonable
to German-speakers?
> integration to the reference frame for astrological coordinates (true
> equinox of date), all corrections like relativistic aberration,
> deflection of light in the gravity field of the Sun etc. have been
> performed with utmost care and precision so that the target precision
> of 0.001 arcsec is maintained through all transformation steps. Never
> before has such a high precision ephemeris been available to
> astrologers.
>
> I think historic details should go to some place in the documentation. The
> main information is purpose, origin and precision.
Given that there's a -doc package this is ludicrous.
>
> Swiss Ephemeris contains three ephemerides. The user can choose
> whether he/she wants to use the original JPL DE406 data (if available
Clunky gender-neutrality; just make them plural in the first place.
> at his/her site), the compressed Swiss Ephemeris data (the default) or
> a built in semianalytic theory by Steve Moshier. The Swiss Ephemeris
^
Missing hyphen.
> package switches automatically to the available best precision
"Best available", not "available best".
> ephemeris dependent on which installed ephemeris files it finds. Even
> without any stored ephemeris files, using the Moshier model, planetary
Wait, do I get to choose the Moshier version or does it automatically
pick the highest precision available? And if I've installed this
package, how would there not be a stored ephemeris?
> positions with better than 0.1 seconds of arc precision are available
> (3 arcsec for the Moon).
>
> At present I don't have good ideas how to shorten this text
>
> In addition to the astronomical planets as contained in the JPL
> integration, Astrodienst's have included all other bodies and
^
Another astrostropher's apology.
> hypothetical factors which are of interest to the astrologer.
^^
I really don't believe they've taken into account every single
hypothetical factor Mystic Meg might care about.
> Astrodienst have used Astrodienst's own numerical integration program
^^^^
I hope they had permission. Oh, and "team plural" - idiomatic for me,
but frowned on in en_US.
> to provide ephemerides for ALL known asteroids. There are over 55'000
^
$LC_NUMERIC. Surely there are at the very least (depending on where
you stop bothering to count) millions of asteroids? 55,000 would be
about right for asteroids bigger than six or seven km diameter, or
perhaps for how many we knew about at some point in the nineties...
> of them and nobody will be able to use them all. Astrodienst distribute
> these extended asteroid files via Astrodienst's download area; there
> are also CDROMs available with large sets of asteroid files. Asteroid
So none of this is relevant to what's in the package.
> reaserachers may be interested in a December 1998 article in the
^^^^^^^^ ^^
Are those like researchers?
> Economist magazine about the naming of asteroids.
It's got to be more interesting than this package description.
>
> Speed: The Swiss Ephemeris is precise and fast. On Astrodienst's Linux
> test machine, a 1000 MHz Pentium III, Astrodienst compute 10'000 complete
^
Never boast about your hardware on a web page you aren't planning to
keep up-to-date. And now I'm sick of the word Astrodienst, too.
> sets of planetary positions, i.e. 10'000 x 11 planets, in 9 seconds. This
^
I suppose that count of planets isn't bad as a rough guess.
> is 0.9 milliseconds for the complete set of exact planetary positions
> (consecutive 1 day steps).
>
> The last paragraph should be skipped: There is too little information to
> compare the machine wis current ones. The processor is quite old, memory
> and OS are not mentioned.
>
>
> Here is my proposol for a shortened version.
>
>
> Description: C library for trajectories of celestial bodies (Swiss Ephemeris)
Too long; I would be willing to put up with something shorter than the
original.
> An ephemeris is a table of values that gives the positions of astronomical
> objects at a given time or times. The Swiss Ephemeris is based upon the
> planetary and lunar ephemeris, DE405/406, developed by NASA's Jet
> Propulsion Laboratory. libswe0 enables programs to access the Swiss
> Ephemeris.
Crunching:
An ephemeris is a table giving astronomical bodies' positions at given
times. The Swiss Ephemeris library gives access to one based on the
DE405/406 ephemeris developed by NASA's Jet Propulsion Laboratory.
> .
> The original integration by the Jet Propulsion Laboritory has been
> extended by Astrodienst, carefully calculating all corrections such as
> relativistic abberation, thus maintainig the target precision of 0.001
> arcsec maintained through all transformation steps. Now it covers the
> years 5400 BC to 5400 AD.
Typos (Laboratory, aberration, maintaining). But it flattens right
down to:
Astrodienst has carefully extended the timeframe to cover the years
5400 BC to 5400 AD while maintaining milliarcsecond precision.
> .
> Swiss Ephemeris contains three ephemerides. The user can choose
> whether he/she wants to use the original JPL DE406 data (if available
> at his/her site), the compressed Swiss Ephemeris data (the default) or
> a built in semianalytic theory by Steve Moshier. The Swiss Ephemeris
> package switches automatically to the available best precision
> ephemeris dependent on which installed ephemeris files it finds. Even
> without any stored ephemeris files, using the Moshier model, planetary
> positions with better than 0.1 seconds of arc precision are available
> (3 arcsec for the Moon).
>
> Perhaps the last sentence could be skipped?
I suggest the entire paragraph could be shifted to a README; its
relevance depends on exactly what data the package contains. Or in
fact isn't it the -data package(s) that would contain these files?
> .
> In addition Astrodienst has included all other bodies and hypothetical
> factors which are of interest to the astrologer. Astrodienst also offers
> ephemerides for ALL known asteroids (over 55'000).
>
> Here I don't understand why the data is targeted at astrologers.
Well, look at astro.com's front page.
Drop the advert for other things not in this package and merge it into
the previous paragraph.
> Please consider adapting the other descriptions as well:
>
> libswe-dev:
>
> Description: C library for trajectories of celestial bodies (development files)
> This package contains static library, headers, example code and
> development manpages for libswe0.
> .
> $some_boilerplate …
>
> swe-basic-data:
>
> Description: C library for trajectories of celestial bodies (basic data files)
> This package includes basic data files needed by libswe, the Swiss
> Ephemeris. The Swiss Ephemeris library can be used without installed data,
> if the user provides that data in her own private directory and points to
> it with SE_EPHE_PATH.
"Her"?
>
> libswe-doc:
>
> Description: C library for trajectories of celestial bodies (documentation)
My own suggestions:
libswe0:
Description: Swiss Ephemeris - library
An ephemeris is a table giving astronomical bodies' positions at given
times. The Swiss Ephemeris library gives access to [...] based on
the DE405/406 ephemeris developed by NASA's Jet Propulsion Laboratory.
.
Astrodienst has carefully extended the timeframe to cover the years
5400 BC to 5400 AD while maintaining milliarcsecond precision, and has
included all other bodies and hypothetical factors of interest to
astrologers.
libswe-dev:
Description: Swiss Ephemeris - development files
An ephemeris is a table giving astronomical bodies' positions at given
times. The Swiss Ephemeris library gives access to one based on the
DE405/406 ephemeris developed by NASA's Jet Propulsion Laboratory.
.
This package contains the static library, headers, example code and
development manpages.
swe-basic-data:
Description: Swiss Ephemeris - basic data files
An ephemeris is a table giving astronomical bodies' positions at given
times. The Swiss Ephemeris library gives access to one based on the
DE405/406 ephemeris developed by NASA's Jet Propulsion Laboratory.
.
This package includes basic data files. However, the library can be
used without packaged data as long as files are available locally in a
directory pointed at by SE_EPHE_PATH.
Should this include a mention of roughly how big the file is? There's
also a swe-standard-data in a separate source package...
libswe-doc:
Description: Swiss Ephemeris - documentation
An ephemeris is a table giving astronomical bodies' positions at given
times. The Swiss Ephemeris library gives access to one based on the
DE405/406 ephemeris developed by NASA's Jet Propulsion Laboratory.
.
This package provides PDF- and HTML-format documentation for users and
developers.
--
JBR with qualifications in linguistics, experience as a Debian
sysadmin, and probably no clue about this particular package
Reply to: