Re: 2.6 Release -- SCheme

[Ted Creedon]

Nothing like progress, installing guile 1.8 doesn't work either
Tedc
On Sunday, January 1, 2012, John Ralls 
wrote:
>
> On Jan 1, 2012, at 7:00 AM, Ted Creedon wrote:
>
>> this talk is all well & good but meanwhile GC is inoperable on updated
>> distros using Guile 2.0.
>
> And that is the distro maintainers problem. Go bug them or switch distros.
>
>>
>> I think one should bite the bullet and eliminate swig/guile.
>>
>
> That isn't in the cards for 2.6, Guile has its tentacles into a lot more
places than the report writer. Until somebody figures out how to get
Guile-2.0 to load extensions from somewhere other than /usr/lib, Guile-2.0
support isn't going to happen either.
>
> Regards,
> John Ralls
>
>
>
___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel

Re: 2.6 Release -- SCheme

[John Ralls]

On Jan 1, 2012, at 7:00 AM, Ted Creedon wrote:

> this talk is all well & good but meanwhile GC is inoperable on updated
> distros using Guile 2.0.

And that is the distro maintainers problem. Go bug them or switch distros.

> 
> I think one should bite the bullet and eliminate swig/guile.
> 

That isn't in the cards for 2.6, Guile has its tentacles into a lot more places 
than the report writer. Until somebody figures out how to get Guile-2.0 to load 
extensions from somewhere other than /usr/lib, Guile-2.0 support isn't going to 
happen either.

Regards,
John Ralls



___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel

Re: 2.6 Release -- SCheme

[Ted Creedon]

this talk is all well & good but meanwhile GC is inoperable on updated
distros using Guile 2.0.

I think one should bite the bullet and eliminate swig/guile.

Tedc

On Sun, Jan 1, 2012 at 4:51 AM, Donald Allen  wrote:

> On Sun, Jan 1, 2012 at 2:39 AM, Gour  wrote:
> > On Sat, 31 Dec 2011 18:36:28 -0500
> > Mike Alexander  wrote:
> >
> >> Python is popular now, but will it be in 10 years?  I've seen
> >> lots of languages come and go.
> >
> > That's true, but I bet it will be more popular than Scheme for sure.
> >
> > Moreover, we can speculate what will happen with GTK+ in 10 years, but
> > let's make GC more approachable *today*. ;)
>
> Well said.
>
> We do things all the time (drive cars, fly in airplanes) that have a
> small, but non-zero, probability of a bad outcome. We just try to make
> smart bets, on a risk/benefit basis. I think choosing Python for this
> application is in that category.
>
> /Don
>
> >
> > Sincerely,
> > Gour
> >
> >
> > --
> > A self-realized man has no purpose to fulfill in the discharge
> > of his prescribed duties, nor has he any reason not to perform
> > such work. Nor has he any need to depend on any other living being.
> >
> > http://atmarama.net | Hlapicina (Croatia) | GPG: 52B5C810
> >
> > ___
> > gnucash-devel mailing list
> > gnucash-devel@gnucash.org
> > https://lists.gnucash.org/mailman/listinfo/gnucash-devel
> >
>
> ___
> gnucash-devel mailing list
> gnucash-devel@gnucash.org
> https://lists.gnucash.org/mailman/listinfo/gnucash-devel
>
___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel

Re: 2.6 Release -- SCheme

[Donald Allen]

On Sun, Jan 1, 2012 at 2:39 AM, Gour  wrote:
> On Sat, 31 Dec 2011 18:36:28 -0500
> Mike Alexander  wrote:
>
>> Python is popular now, but will it be in 10 years?  I've seen
>> lots of languages come and go.
>
> That's true, but I bet it will be more popular than Scheme for sure.
>
> Moreover, we can speculate what will happen with GTK+ in 10 years, but
> let's make GC more approachable *today*. ;)

Well said.

We do things all the time (drive cars, fly in airplanes) that have a
small, but non-zero, probability of a bad outcome. We just try to make
smart bets, on a risk/benefit basis. I think choosing Python for this
application is in that category.

/Don

>
> Sincerely,
> Gour
>
>
> --
> A self-realized man has no purpose to fulfill in the discharge
> of his prescribed duties, nor has he any reason not to perform
> such work. Nor has he any need to depend on any other living being.
>
> http://atmarama.net | Hlapicina (Croatia) | GPG: 52B5C810
>
> ___
> gnucash-devel mailing list
> gnucash-devel@gnucash.org
> https://lists.gnucash.org/mailman/listinfo/gnucash-devel
>

___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel

Re: 2.6 Release -- SCheme

[Gour]

On Sat, 31 Dec 2011 18:36:28 -0500
Mike Alexander  wrote:

> Python is popular now, but will it be in 10 years?  I've seen
> lots of languages come and go.

That's true, but I bet it will be more popular than Scheme for sure.

Moreover, we can speculate what will happen with GTK+ in 10 years, but
let's make GC more approachable *today*. ;)

Sincerely,
Gour


-- 
A self-realized man has no purpose to fulfill in the discharge 
of his prescribed duties, nor has he any reason not to perform 
such work. Nor has he any need to depend on any other living being.

http://atmarama.net | Hlapicina (Croatia) | GPG: 52B5C810


signature.asc
Description: PGP signature
___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel

Re: 2.6 Release -- SCheme

[Gour]

On Sat, 31 Dec 2011 18:31:34 -0500
Donald Allen  wrote:

> If there is agreement among the developers that this is an attractive
> alternative to Guile, I would advise doing some prototyping to
> evaluate the performance of Python for report generation. 

I believe you ('cause I did not dive into GC yet) that performance might
be problem, but in my case customizability is much bigger one, so we're
ready to trade it for performance.

Otoh, it's not we would not believe with Python, but just wonder if you
thought about Lua which shoould go nicely along with C and it is very
simple language to learn for end user wanting to customize reports?


Sincerely,
Gour

-- 
A person who has given up all desires for sense gratification, 
who lives free from desires, who has given up all sense of 
proprietorship and is devoid of false ego — he alone can 
attain real peace.

http://atmarama.net | Hlapicina (Croatia) | GPG: 52B5C810


signature.asc
Description: PGP signature
___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel

Re: 2.6 Release -- SCheme

[Gour]

On Sat, 31 Dec 2011 18:17:13 -0500
"Derek Atkins"  wrote:

> Is it really worth our time to find another scheme implementation and
> swap everything over to it?  I would think that it would be better to
> write a report infrastructure in a language that would seem more
> "popular" (python), build in the infrastructure, and then send out a
> call for report writers to convert the existing scheme reports over
> to the new language.

+1

Very well put together. ;)

btw, wishing happy & prosperous New Year to everybody to make GC even
better in 2012!


Sincerely,
Gour


-- 
Bewildered by the modes of material nature, the ignorant fully 
engage themselves in material activities and become attached. But 
the wise should not unsettle them, although these duties are inferior 
due to the performers' lack of knowledge.

http://atmarama.net | Hlapicina (Croatia) | GPG: 52B5C810


signature.asc
Description: PGP signature
___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel

Re: 2.6 Release -- SCheme

[Donald Allen]

On Sat, Dec 31, 2011 at 6:41 PM, Mike Alexander  wrote:
> --On December 31, 2011 6:31:34 PM -0500 Donald Allen
>  wrote:
>
>> If there is agreement among the developers that this is an attractive
>> alternative to Guile, I would advise doing some prototyping to
>> evaluate the performance of Python for report generation. My guess is
>> that it will be faster than Guile, but that's only a guess. It should
>> be tested before committing to it. One of the problems (and not the
>> only one) with the current report system is performance, and it would
>> not be smart to invest a lot of effort in a new system that had the
>> same problem.



>>
>
> Having looked at a number of reports, and optimized some by factors of 5 or
> more, I don't think changing the language will help this problem. Even if
> the Guile interpreter were infinitely fast, some reports would still be slow
> because of the algorithms they use.

If the Guile interpreter were infinitely fast, we wouldn't be having
this discussion; the reports would be infinitely fast :-)

  They often do multiple passes over
> large searches through the Gnucash data.  That is where the time often goes.
>  Of course if all the reports were rewritten, these problems might be fixed
> in passing.

A bad algorithm written in a fast language running on a fast computer
can be unacceptably slow. A good algorithm written in a slow language
on a fast computer can be unacceptably slow. Yes, there's more
leverage in the algorithm, but the language can matter, too.

/Don

>
>        Mike
>
>
> ___
> gnucash-devel mailing list
> gnucash-devel@gnucash.org
> https://lists.gnucash.org/mailman/listinfo/gnucash-devel

___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel

Re: 2.6 Release -- SCheme

[Mike Alexander]

--On December 31, 2011 6:31:34 PM -0500 Donald Allen 
 wrote:



If there is agreement among the developers that this is an attractive
alternative to Guile, I would advise doing some prototyping to
evaluate the performance of Python for report generation. My guess is
that it will be faster than Guile, but that's only a guess. It should
be tested before committing to it. One of the problems (and not the
only one) with the current report system is performance, and it would
not be smart to invest a lot of effort in a new system that had the
same problem.



Having looked at a number of reports, and optimized some by factors of 
5 or more, I don't think changing the language will help this problem. 
Even if the Guile interpreter were infinitely fast, some reports would 
still be slow because of the algorithms they use.  They often do 
multiple passes over large searches through the Gnucash data.  That is 
where the time often goes.  Of course if all the reports were 
rewritten, these problems might be fixed in passing.


Mike

___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel

Re: 2.6 Release -- SCheme

[Mike Alexander]

--On December 31, 2011 6:17:13 PM -0500 Derek Atkins  
wrote:




On Sat, December 31, 2011 5:46 pm, Mike Alexander wrote:

--On December 31, 2011 6:22:55 PM + Hendrik Boom
 wrote:


On Fri, 30 Dec 2011 18:19:58 +0100, Geert Janssens wrote:

This is probably a more drastic change than guile 2.0, but:

There's another implementation of Scheme available that actually
compiles  Scheme to C or C++ -- Gambit-C.  You can actually embed
C++ code within  the C code, even #include stuff.  There's also an
interpreter, but the  interpreter doesn't have embedded C/C++ code,
though it can call  previously compiled code that does.  The Debian
package is called gambc. I have no idea whether this would be easier
to use and maintain than  using guile.


I don't know anything about Gambit-C, but it's also available in
MacPorts on MacOSX.  The description sounds promising.  The home page
is at

where they also link to Windows (and iPad!) installers.


Is it really worth our time to find another scheme implementation and
swap everything over to it?  I would think that it would be better to
write a report infrastructure in a language that would seem more
"popular" (python), build in the infrastructure, and then send out a
call for report writers to convert the existing scheme reports over
to the new language.


That's a very good question.  I think the answer depends to some extent 
on the effort involved.  Switching languages will likely be 
non-trivial, switching to Gambit might be easier (I don't really know). 
I wasn't advocating switching to Gambit, just pointing out that this is 
a possibility that might be worth considering.  As I said I know almost 
nothing about Gambit.  Switching languages might be better but choosing 
a language may be more difficult than it seems.  Python is popular now, 
but will it be in 10 years?  I've seen lots of languages come and go.


 Mike

___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel

Re: 2.6 Release -- SCheme

[Donald Allen]

On Sat, Dec 31, 2011 at 6:17 PM, Derek Atkins  wrote:
>
> On Sat, December 31, 2011 5:46 pm, Mike Alexander wrote:
>> --On December 31, 2011 6:22:55 PM + Hendrik Boom
>>  wrote:
>>
>>> On Fri, 30 Dec 2011 18:19:58 +0100, Geert Janssens wrote:
>>>
>>> This is probably a more drastic change than guile 2.0, but:
>>>
>>> There's another implementation of Scheme available that actually
>>> compiles  Scheme to C or C++ -- Gambit-C.  You can actually embed C++
>>> code within  the C code, even #include stuff.  There's also an
>>> interpreter, but the  interpreter doesn't have embedded C/C++ code,
>>> though it can call  previously compiled code that does.  The Debian
>>> package is called gambc. I have no idea whether this would be easier
>>> to use and maintain than  using guile.
>>
>> I don't know anything about Gambit-C, but it's also available in
>> MacPorts on MacOSX.  The description sounds promising.  The home page
>> is at 
>> where they also link to Windows (and iPad!) installers.
>
> Is it really worth our time to find another scheme implementation and swap
> everything over to it?  I would think that it would be better to write a
> report infrastructure in a language that would seem more "popular"
> (python), build in the infrastructure, and then send out a call for report
> writers to convert the existing scheme reports over to the new language.

Much as I prefer working in Scheme to any other language for most
things, I was about to write the same message. Python is very usable,
well-supported and well-documented. And it's far better accepted in
the general community than Scheme. I agree completely with Derek.

If there is agreement among the developers that this is an attractive
alternative to Guile, I would advise doing some prototyping to
evaluate the performance of Python for report generation. My guess is
that it will be faster than Guile, but that's only a guess. It should
be tested before committing to it. One of the problems (and not the
only one) with the current report system is performance, and it would
not be smart to invest a lot of effort in a new system that had the
same problem.

/Don

>
>>          Mike
>
> -derek
> --
>       Derek Atkins                 617-623-3745
>       de...@ihtfp.com             www.ihtfp.com
>       Computer and Internet Security Consultant
>
> ___
> gnucash-devel mailing list
> gnucash-devel@gnucash.org
> https://lists.gnucash.org/mailman/listinfo/gnucash-devel

___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel

Re: 2.6 Release -- SCheme

[Derek Atkins]

On Sat, December 31, 2011 5:46 pm, Mike Alexander wrote:
> --On December 31, 2011 6:22:55 PM + Hendrik Boom
>  wrote:
>
>> On Fri, 30 Dec 2011 18:19:58 +0100, Geert Janssens wrote:
>>
>> This is probably a more drastic change than guile 2.0, but:
>>
>> There's another implementation of Scheme available that actually
>> compiles  Scheme to C or C++ -- Gambit-C.  You can actually embed C++
>> code within  the C code, even #include stuff.  There's also an
>> interpreter, but the  interpreter doesn't have embedded C/C++ code,
>> though it can call  previously compiled code that does.  The Debian
>> package is called gambc. I have no idea whether this would be easier
>> to use and maintain than  using guile.
>
> I don't know anything about Gambit-C, but it's also available in
> MacPorts on MacOSX.  The description sounds promising.  The home page
> is at 
> where they also link to Windows (and iPad!) installers.

Is it really worth our time to find another scheme implementation and swap
everything over to it?  I would think that it would be better to write a
report infrastructure in a language that would seem more "popular"
(python), build in the infrastructure, and then send out a call for report
writers to convert the existing scheme reports over to the new language.

>  Mike

-derek
-- 
   Derek Atkins 617-623-3745
   de...@ihtfp.com www.ihtfp.com
   Computer and Internet Security Consultant

___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel

Re: 2.6 Release -- SCheme

[Mike Alexander]

--On December 31, 2011 6:22:55 PM + Hendrik Boom 
 wrote:



On Fri, 30 Dec 2011 18:19:58 +0100, Geert Janssens wrote:

This is probably a more drastic change than guile 2.0, but:

There's another implementation of Scheme available that actually
compiles  Scheme to C or C++ -- Gambit-C.  You can actually embed C++
code within  the C code, even #include stuff.  There's also an
interpreter, but the  interpreter doesn't have embedded C/C++ code,
though it can call  previously compiled code that does.  The Debian
package is called gambc. I have no idea whether this would be easier
to use and maintain than  using guile.


I don't know anything about Gambit-C, but it's also available in 
MacPorts on MacOSX.  The description sounds promising.  The home page 
is at  
where they also link to Windows (and iPad!) installers.


Mike

___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel

Re: 2.6 Release -- SCheme

[Hendrik Boom]

On Fri, 30 Dec 2011 18:19:58 +0100, Geert Janssens wrote:

> Op vrijdag 30 december 2011 09:06:58 schreef u:
> 
>> Swig/Guile: It looks to me like we have a much broader problem: Swig's
>> Guile support is not maintained. For the short term we can try applying
>> the patch from the Swig bug report and see if that gets us Guile 2.0
>> support, but in the longer run it looks like we need to either switch
>> back to GWrap or replace Guile with something that's better supported.
>> 
> Yes, this is a bad problem.

This is probably a more drastic change than guile 2.0, but:

There's another implementation of Scheme available that actually compiles 
Scheme to C or C++ -- Gambit-C.  You can actually embed C++ code within 
the C code, even #include stuff.  There's also an interpreter, but the 
interpreter doesn't have embedded C/C++ code, though it can call 
previously compiled code that does.  The Debian package is called gambc.
I have no idea whether this would be easier to use and maintain than 
using guile.

-- hendrik

___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel