I run the risk that this post is being reported for advertising
Thursday afternoon I finished a series of meaurements for which the report should be finished and on the desk of my client by Friday 17:00. And I had to fly home Thursday afternoon as well.
What I had was a number of CSV files with inside those files buried the measurement data.
The report I had to produce would contain 20 of those graphs, a bunch of JPG images and about 5 tables. The report had to be "scientific", so table of contents, header, footer, correct figure, lemma and table numbering, section numbering, cross references, bibliography, the works.
I did have 80% of all this formats in previous reports, so I did not have to re-invent the wheel, and it was not my first Latex report either.
Nevertheless, producing these graphs from the CSV files
was new, but went fast thanks to some Gnuplot scripts. Including titles, scales and annotation. As a matter of fact I only had to write two versions of the graphs and the rest was copy & paste.
All the knowledge of the report was in my head, but it had to be ordered, divided in chapters and set up logically.
Eventually the report contained 29 pages. Including proofreading it was finished on Friday at 16:45.
Until Friday 16:00 I spent
all my time composing the report, wording, correct set-up clear wording of conclusions etc. In vim. At 16:00 I started proofreading, which is surprisingly easy because the layout is so different that the author hardly suffers from text blindness which usually occurs when reading your own documents. In my experience, proofeading the PDF once is sufficient to get out 95% of all errors.
The next 45 minutes I spent correcting logical, style and grammatical errors. Time spent for layout, correction of numbering, captions and references: 30 seconds.
This is incredible.
Everyone who has ever written a serious report in Word or Writer knows that the author has to spend hours and hours in correcting styles, layout, captions, references, paragraph numbering, table of contents, and then correcting over and over again. Graphs tend to jump to locations where the don't belong, to the footer, or to page 1, but not where they should be. Captions disappear or appear twice, whatever. The more the deadline approaches, the worse it gets. The auther gets completely text blind as he has read thru the text a zillion times already. The worst thing is, time to produce does not decrease with increasing experience of the author
as the same actions error corrections have to be repeated over and over again.
I can't remember I have ever written a report this fast.
jlinkels