3 Jun
2024
3 Jun
'24
12:21 a.m.
On Sun, Jun 2, 2024 at 1:08 AM Marc Rochkind <mrochkind(a)gmail.com> wrote:
True enough, Kevin, but with the decline of printed books and the increase in online
docs, I rarely find what I'm looking for in a printed book and, when I think I have,
the price is very high for what may turn out to be a bad guess. Browsing a bookstore for
serious computer books is no longer possible, except maybe in very large cities.
Agreed, bookstores are more or less dead. I used the Internet Archive
a lot to inform my pre-purchasing decisions but the copyright
enforcement has caught up there.
For example, for an upcoming project I need up-to-date
and authoritative information on Kotlin and AWS S3 APIs.
I believe there are decent Kotlin books out. There are some "fast"
publishers like Manning, Apress, and Packt (maybe in rough order of
quality..) that put out a lot of ephemeral literature but occasionally
have some fairly good works. There aren't a lot of consistent bangers
like Prentice-Hall PTR was putting out back in the day although I am
generally impressed with some of the work Pearson is putting out. No
Starch is also generally a winner, although a little less hard sciency
and more pop.
S3 is, as a user, so trivial I am not sure it warrants a book. In the
past "cookbook" style books were common and maybe even useful. When I
was getting started, I was thirsty for easy copy+paste solutions so
that I didn't have to strain much thought to get results. I believe
Large Language Models are good enough to subsume some of that now.
On the other hand, a good book on building applications in a
cloud-native way definitely will shave a year or two off the learning
curve. What and why seem to be more enduring than how.
Living in the past, I find, is no help!
I don't think I live in the past, I am working on similar technologies
you mention to earn a living in the present. One thing I failed to
mention in my post, and I think related to all this is the utility of
Large Language Models. In your example above, the best current LLMs
would be helpful for S3 and a little less so (but not useless) for
Kotlin. However, LLMs still can't really help with the synthesis of
good overall design and taste while an enduring book will impart both
on an intrepid reader that should outlive the details being discussed.
No doubt, whatever you are doing now is informed by your past.
One other anecdote, in my recent passion of learning digital logic
design, I find even the most recent textbooks are well referenced to
papers and books of the past which is a bit of a contrast to
programming literature. Most will go back to Boole's "Studies of
Logic & Probability" as the basis. Lots of papers referenced from the
40s and books from the 70s and 80s still have authority if you are
serious about the subject - Quine, McClusky, RK Richards, etc had a
lot to say early on and it is very much still valid.
Marc Rochkind
(author of the first book on UNIX programming)
Yes, I recognized your name and have your books.
On Sun, Jun 2, 2024, 7:12 AM Kevin Bowling <kevin.bowling(a)kev009.com> wrote:
>
> On Sat, Jun 1, 2024 at 7:31 PM Will Senn <will.senn(a)gmail.com> wrote:
> >
> > Today, as I was digging more into nroff/troff and such, and bemoaning the lack
of brevity of modern text. I got to thinking about the old days and what might have gone
wrong with book production that got us where we are today.
> >
> > First, I wanna ask, tongue in cheek, sort of... As the inventors and early
pioneers in the area of moving from typesetters to print on demand... do you feel a bit
like the Manhattan project - did you maybe put too much power into the hands of folks who
probably shouldn't have that power?
> >
> > But seriously, I know the period of time where we went from hot metal
typesetting to the digital era was an eyeblink in history but do y'all recall how it
went down? Were you surprised when folks settled on word processors in favor of markup? Do
you think we've progressed in the area of ease of creating documentation and printing
it making it viewable and accurate since 1980?
> >
> > I didn't specifically mention unix, but unix history is forever bound to
the evolution of documents and printing, so I figure it's fair game for TUHS and
isn't yet COFF :).
> >
> > Later,
> >
> > Will
>
> I think your other topic is closely related but I chose this one to reply to.
>
> I own something well north of 10,000 technical and engineering books
> so I will appoint myself as an amateur librarian.
>
> When I was younger, I had the false notion that anything new is good.
> This attitude permates a lot of society. Including professional
> libraries. They have a lot of collection management practices around
> deciding what and when to pitch something and a big one is whether the
> work is still in print, while a more sophisticated collection will
> also take into account circulation numbers (how often it is checked
> out). A lot of that is undoubtedly the real costs surrounding storing
> and displaying something (an archived book has a marginal cost, a
> publically accessible displayed book presumably has a higher
> associated cost) as well as the desire to remain current and provide
> value to the library's membership.
>
> From what I have seen, there isn't much notion of retaining or
> promoting a particular work unless it remains in print. As an
> example, K&R C is still in print and would be retained by most
> libraries. The whole thing becomes a bit ouroboros because that leads
> to more copies being printed, and it remaining in collections, and
> being read. Obviously, this is a case of a great piece of work
> benefiting from the whole ordeal. But for more niche topics, that
> kind of feedback loop doesn't happen. So the whole thing comes down
> in a house of cards... the publisher guesses how many books to print,
> a glut of them are produced, they enter circulation, and then it goes
> out of print in a few years. A few years later it is purged from the
> public libraries. As an end user, one benefit to this collapse is
> that used books are basically flooded into the market and you can get
> many books for a fraction of their retail price used.. but it becomes
> difficult to know _what_ to get if you don't have an expert guide or
> somewhere to browse and select for yourself.
>
> So why does this all matter to your more meta question of why less
> great books? There is less to no money in it nowadays for authors.
> The above example of library circulation represented a large number of
> guaranteed sales to wealthy institutions (academic and government =
> wealth, don't let them pretend otherwise). Except now many libraries
> have downsized their physical collections to make room for multimedia
> or just lower density use of space. So there are less guaranteed
> sales.
>
> Another facet of the same coin, one reason printed books are great has
> to do with the team surrounding their production. If you look near
> the colophon, you will often find a textbook will have quite a few
> people involved in moving a manuscript to production. This obviously
> costs a lot of money. As things move more to ebook and print on
> demand, it's an obvious place to cut publishing expenses and throw all
> the work directly onto the author. That may result in cheaper books
> and maybe(?) more revenue for the author, but it won't have the same
> quality that a professional publishing team can bring to the table.
>
> As to my deliberate decision to accumulate the dead trees and ink,
> it's because although online docs are great I find my best learning is
> offline while I use the online docs more like mental jogs for a
> particular API or refamiliarizing myself with the problem domain. I
> have some grandeur ambitions that first involve a large scanning
> project but that will have to await more self funding.
>
> Regards,
> Kevin