2010/8/29 Sudhakar Krishnamachari
<skrishnamach...@gmail.com>
My two cents long time in my blog on exactly the same subject:
Its great to see a free version of Smalltalk at last measuring up to the
needs of using it in an business sense and showing it to non-Smalltalker=
=92s
and the =93IT managers=94 and not put them off. I am sure this is going t=
o make
it difficult for VA and Cincom to create new small-medium shop sales goin=
g
forward especially with Seaside, Magritte and raft of other frameworks
actually being born in Squeak. There obviously is the crying need for
High Priority:
Enusure very high reliability of the base Pharo. No bugs that are not
fixable by the average developer with a bit of google and sending a few
emails. If it needs that one expert in one other end.. business continuit=
y
is hugely impacted
Primary:
* better-faster VM aka Cog mainstream
* Highly performant DBMS connectivity natively
Hi and thanks for the notes.
There is NO way to have native Smalltlak driver for propietary databases.
Even if I would love to have the database driver implemented in Smalltalk
rather than calling a C library, this is not possible with propiertary
databases. If you don't have a public protocol or API there is no way to
build the driver. Do you think Oracle or Microsoft would do a Smalltalk
driver? I doubt it. So....for propietary databases....you have to somehow,
use of the the existing database drivers.
ALL databases have a driver in C, usually called Database Client Library. C
is fast, and we have FFI. That's why SqueakDBX choosed OpenDBX.
akin to JDBC through to all DBs no DB specific stuff like now, the Squeak
DBX is not business friendly.
Are you talking about SqueakDBX project ? (http://www.squeakdbx.org/)
If so, I would like you say it is not businness friendly.
* Comprehensive Web Services + XMLRPC
* Pharo-Groovy bridge and / or Pharo-Ruby bridge ( or can we make Pharo
work in tandem with a JVM..! this will help leverage the tons of framewo=
rks
in existence even if they are tad slower than native Java)
* Configuration Management System plugin for SVN
There is SqueakSVN project. And I think in SCM they are doing to GIT.
* Acceleration of easy docs and plugins/ web development related framewor=
k
interfaces from Seaside viz flex/ silverlight: many that are possible as =
of
now, but needs a Smalltalk expert to cobble it up
* Have not really gone deeper with Webclient/ FTPClient/ SMTPClient et al=
s.
they need to be really simple/ highly reliable/ dumbed down for the avera=
ge
coder to pick snippets and make it work
May be lots more.. I can bring in lot more thought this in terms of
Generic/ Sectoral or Domain/ Type of Apps preferences viz:
Banking/Insurance/Utilities/Technical/ Web vs GUI etc ...So first we shou=
ld
think hard on the segment you want to hit harder..
Secondary:
* Better tools for GUI Dev and more standard GUI development capabilities
* Make Pharo easily usable for a team of 100+ developers together in
one/multiple locations with code synch issues, change / version control ,
performance ( a host of other stuff included) taken care of.
Agree. Torch project seems that could help here. Altoguht Monticello 2 or
similar would be needed.
* Pluggable to industry standard tools for development: Rational/ Togethe=
r
Soft viz for UML creation and conversion to code / reverse engineer code =
to
UML
Carla worked on UML for this year GSoC project. It would probably be
integrated soon in Pharo.
Follow through with:
* Pluggable with QTP/ Testing Automation tools.. Actually this is very
important in the standard large volume development. Makes a huge differen=
ce
for the risk mitigation for the company. really inexpensive , kind of not=
so
highly technical staff runs this show in comparison to Unit Tests/ test
automation through ST code is a more riskier investment and I have not se=
en
in JAva/ .Net / or even Smalltalk large legacy projects unit test done an=
d
maintained well enough
-Skrish
_______________________________________________
Pharo-project mailing list
Pharo-project@lists.gforge.inria.fr
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
<br><br><div class=3D"gmail_quote">2010/8/29 Sudhakar Krishnamachari <span =
dir=3D"ltr"><<a
href=3D"mailto:skrishnamach...@gmail.com";>skrishnamachar=
class=3D"gmail_quote" style=3D"ma=
rgin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding=
-left: 1ex;">
My two cents long time in my blog on exactly the same subject:<br><br>Its g=
reat to see a free version of Smalltalk at last measuring up to
the needs of using it in an business sense and showing it to
non-Smalltalker=92s and the =93IT managers=94 and not put them off. I am su=
re
this is going to make it difficult for VA and Cincom to create new
small-medium shop sales going forward especially with Seaside, Magritte
and raft of other frameworks actually being born in Squeak. There
obviously is the crying need for<br><br>High Priority:<br>Enusure very high=
reliability of the base Pharo. No bugs that are not fixable by the average=
developer with a bit of google and sending a few emails. If it needs that =
one expert in one other end.. business continuity is hugely impacted<br>
<br>Primary:<br>* better-faster VM aka Cog mainstream<br>* Highly performan=
t DBMS connectivity
natively</blockquote><div><br>Hi and thanks for the notes. <br>There is NO =
way to have native Smalltlak driver for propietary databases. Even if I wou=
ld love to have the database driver implemented in Smalltalk rather than ca=
lling a C library, this is not possible with propiertary databases. If you =
don't have a public protocol or API there is no way to build the driver=
. Do you think Oracle or Microsoft would do a Smalltalk driver? I doubt it.=
=A0 So....for propietary databases....you have to somehow, use of the the e=
xisting database drivers. <br>
<br>ALL databases have a driver in C, usually called Database Client Librar=
y. C is fast, and we have FFI. That's why SqueakDBX choosed OpenDBX. <b=
r>=A0</div><blockquote class=3D"gmail_quote" style=3D"margin: 0pt 0pt 0pt 0=
.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
akin to JDBC through to all DBs no DB specific stuff like now, the Squeak =
DBX is not business friendly.<br></blockquote><div><br>Are you talking abou=
t SqueakDBX project ?=A0 (<a href=3D"http://www.squeakdbx.org/";>http://www.=
squeakdbx.org/</a>)<br>
If so, I would like you say it is not businness friendly. <br><br>=A0</div>=
<blockquote class=3D"gmail_quote" style=3D"margin: 0pt 0pt 0pt 0.8ex; borde=
r-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">* Comprehensive W=
eb Services + XMLRPC<br>
* Pharo-Groovy bridge and / or Pharo-Ruby bridge ( or can we make Pharo wor=
k in tandem with a JVM..!=A0 this will help leverage the tons of frameworks=
in existence even if they are tad slower than native Java)<br>
* Configuration Management System plugin for SVN<br></blockquote><div><br>T=
here is SqueakSVN project. And I think in SCM they are doing to GIT. <br>=
=A0</div><blockquote class=3D"gmail_quote" style=3D"margin: 0pt 0pt 0pt 0.8=
ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
* Acceleration of easy docs and plugins/ web development related framework =
interfaces from Seaside viz flex/ silverlight: many that are possible as of=
now, but needs a Smalltalk expert to cobble it up<br>
* Have not really gone deeper with Webclient/ FTPClient/ SMTPClient et als.=
they need to be really simple/ highly reliable/ dumbed down for the averag=
e coder to pick snippets and make it work<br><br>May be lots more.. I can b=
ring in lot more thought this in terms of Generic/ Sectoral or Domain/ Type=
of Apps preferences viz: Banking/Insurance/Utilities/Technical/ Web vs GUI=
etc ...So first we should think hard on the segment you want to hit harder=
.. <br>
<br>Secondary:<br>* Better tools for GUI Dev and more standard GUI developm=
ent capabilities<br>* Make Pharo easily usable for a team of 100+ developer=
s together in one/multiple locations with code synch issues, change / versi=
on control , performance ( a host of other stuff included) taken care of. <=
br>
</blockquote><div><br>Agree. Torch project seems that could help here. Alto=
guht Monticello 2 or similar would be needed.<br>=A0</div><blockquote class=
=3D"gmail_quote" style=3D"margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid=
rgb(204, 204, 204); padding-left: 1ex;">
* Pluggable to industry standard tools for development: Rational/ Together =
Soft viz for UML creation and conversion to code / reverse engineer code to=
UML<br></blockquote><div><br><br>Carla worked on UML for this year GSoC pr=
oject. It would probably be integrated soon in Pharo.<br>
=A0</div><blockquote class=3D"gmail_quote" style=3D"margin: 0pt 0pt 0pt 0.8=
ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;"><br>Foll=
ow through with:<br>* Pluggable with QTP/ Testing Automation tools.. Actual=
ly this is very important in the standard large volume development. Makes a=
huge difference for the risk mitigation for the company. really inexpensiv=
e , kind of not so highly technical staff runs this show in comparison to U=
nit Tests/ test automation through ST code is a more riskier investment and=
I have not seen in JAva/ .Net / or even Smalltalk large legacy projects un=
it test done and maintained well enough<br>
<br><br>-Skrish<br><br><br>
<br>_______________________________________________<br>
Pharo-project mailing list<br>
<a
href=3D"mailto:Pharo-project@lists.gforge.inria.fr";>pharo-proj...@lists.=
