PDF text extraction?

RM richmondmathewson at gmail.com
Sat Apr 2 12:50:58 EDT 2016



On 2.04.2016 19:17, Richard Gaskin wrote:
> Richmond wrote:
>
> > On 2.04.2016 18:38, Richard Gaskin wrote:
> >> Thanks. Yes, Mark Lucas has been doing some outstanding work on
> >> SuperCard 4.8.
> >
> > Well, outstanding is as outstanding does, I really wonder how
> > Supercard keeps going in the face of competition of Livecode.
> > I know that Supercard has been around for donkey's ages (recall
> > playing with it [and finding it rather awkward compared with
> > Hypercard 2.4.1] about 20 years ago), but as Macintosh, whichever
> > way one looks at things, is a coterie, niche market, a multiplatform
> > alternative (pace Livecode) would seem to make it redundant.
>
> One could equally ask about LiveCode in the face of competition from 
> the nearly-ubiquitous Python.  There are so many languages in the 
> world because each offers something a little different from the others.

I compared Supercard and Livecode because, I at least, see them as 
belonging to a family of languages that include things like Toolbook as 
well, to which
Python does not belong: a family of programming languages with an 
integrated compiler and a WYSYWIG GUI.

>
> Given the similarities between LC and SC, it might seem at first 
> glance like they do the same thing, and in some respects they do.
>
> But when you spend more time with them, it becomes clear that the 
> similarities between them are like the similarities between Mac and 
> Windows:  both have overlapping windows and many other nearly 
> identical GUI elements, but Mac is built by a company that also makes 
> the hardware so they're able to exercise a bit more control over the 
> experience, one which its fans feel is more integrated and satisfying.

What that really means is that the Livecode people have to work quite a 
lot harder feature-per-feature than the Supercard people.

>
> Similarly, SuperCard has 1/7th of the job that LC has: OS X vs OS X, 
> Windows, Linux, iOS, Android, Linux/ARM, and faceless Server. Not 
> surprisingly, focused exclusively on 1/7th of the platform coverage, 
> Mark Lucas has been able to integrate with OS X more smoothly in some 
> ways than LC can.
>
> SuperCard is a glove wrapped around the hand of OS X, while LC is more 
> of a platform-independent VM, interfacing with the host OSes in such a 
> broad range of ways that it has to use more internal code for things 
> SC can depend on OS X for.
>
> They look similar because both do a good job, but under the hood their 
> architectures are vastly different, each comprised of good choices but 
> different because they serve very different goals.

The sabre-toothed tiger that roamed across Europe, and the sabre-toothed 
marsupial tiger that roamed across South America had
vastly different interior workings, even if, superficially, they looked 
almost identical.

Neither would have survived long in the other's habitat.

>
>
> > In a perfect world (which is a silly turn of phrase) Livecode would
> > be able to parse just about any file one could chuck at it.
>
> That's the world I hope we're moving toward.  And format by format 
> we'll get there.
>
> Right now most data scientists use Python as their glue language to 
> wrangle Hadoop, R, and the other parts of their workflow.

Data handling is not Livecode's greatest strength at the moment: I think 
it is, despite some naysayyers, still the Hypercard "thing".

If Livecode wants to be a man for all seasons then there are quite a few 
areas which need close attention.  I have sufficient confidence
in Livecode to believe that their team is capable of that (or, if they 
are not, at least seeing that and hiring extra people who are capable of 
that).

> There's really no reason they couldn't be using LiveCode in that role, 
> and it becomes a compellingly self-evident choice once we can move 
> some of that workflow directly into the language through automated 
> handling of more formats.
>

I have children who attend my language school who cannot understand why 
the High schools in this country teach teenagers programming
using either PASCAL (!) or C++ when  Livecode produces results far 
faster while not dumbing things down to the extent that one ends up with
a LEGO-kit.

The reasons in Bulgaria are both historical and due to the low-pay = 
low-motivation-to-change given to school teachers.

This summer I am going to have an "innovative" class consisting of 14-16 
year olds who have been studying C++, and we are going
to duplicate, over a 6 week period (2 or 3 times a week) all the 
projects they have had to work on in 2 years. I intend to record these 
projects
(probably write up the projects as they are presented in the kids' 
textbooks) as a sort of "nyaah, nyahh, nyaah" to the extremely stuffy 
and conservative
state educational system here. Ideally this will end up as a sort of 
monograph; both in English and Bulgarian.

Richmond.

Richmond.




More information about the use-livecode mailing list