SourceForge.Net 2007 Community Choice Awards + O'Reilly OSCON Open Source Awards 2007


2007-07-31 04:57:46 PM
delphi69
See at
sourceforge.net/community/index.php/landing-pages/cca07/
radar.oreilly.com/archives/2007/07/oscon_open_sour_1.html
Feel free to comment!
I'll try to do a small start here saying some words from a Delphian personal POV:
Best Tool or Utility for Developers: Winner: TortoiseSVN
- Delphi provides/provided some tools for team integration, but not to much, imho.
What is needed, imho, isn't so much tools for this, but rather to prepare the _files_
for an easy Check-in/Check-out, and, more generally, to allow an easy integration of
(almost) any team development tool on top of Delphi's IT environment (ie. properly
formatted files, IDE architecture aso).
For me is quite significant that a team tool was the winer.
Best Project for the Enterprise: Winner: Firebird
- Obvious. Imho, its a triumph of functionality and ease of use - Firebird just works
in a natural way, doing what it is supposed to do in the way in which it is supposed to
do it (well, in most of the cases ;-)) - rather than the bunch of features in a
feature matrix. Also, imho, one of the most important success factors for Delphi
was/is that it allows you to concentrate on your _human_ problem (_What_ to do), not
on the code (_how_ to do it). Same stands for Firebird. Now, what it is needed to do
from Delphi POV? Obvious (again). See QC. In fact, one of the most important things
which are missing to Delphi to became a Db powerhouse is, imho, the DBX drivers for
the 'three musketeers': Firebird, PostgreSQL and SQLite. And this isn't at all hard
to do... only a small amount of good willing... ;-)
Best Project for Gamers: Winner: ScummVM
- Whoah!! The winner is a programming language for games. Well, I am not interested in
games, but, imho, it is very significant that a programming language wins in front of
some 3D rendering engines which can be regarded as quite powerful. Imho, one of the
clear needs today (and hence tendencies) is the need, both in business applications
and in {*word*143}/multimedia the need of flexibility and quick (user) configuration which
a scripting language can provide, if it is integrated with a compiled backend. I don't
suggest to do a Scumm implementation for Delphi, nor a Lua one (even Lua is much more
popular and general than Scumm) but a runtime compiler with a Pascal syntax weighting
200-300 kb wrapped in a component with near-to seamless integration with the code
(ie. two-way communication (parameters, object & class usage), syntax check at
compile-time, compiled code reuse, compiling from a stream aso.) - this will be a a
_very_ nice thing to have.
Best Project for Multimedia: Winner: Audacity
- As Delphi user only two hints:
1. Put also a basic support for DirectShow not only of MCI (as TMediaPlayer has)
It's a shame that we need to use 3rd party solutions just to display a movie in an
application.
0. (!!!) Support _all_ the graphics formats which really counts in a optimized,
general-purpose graphics engine (Graphics32.pas for ex.). Nowadays these formats are
very few: BMP (already supported), JPEG (already supported), PNG (GraphicEx - needed
also for Vista), GIF (Where is it? - I could not find it in my Delphi 2007 SKU...), TIFF
(already exists a free implementation for TiffLib, besides what GraphicEx has to
offer - it is de facto standard in technical imaging) and JPEG2000 (this is a little
bit more difficult). that is all. With a small amount of work, we'll have a solid base
for the entire range of CMS/DMS applications/integrated modules.
2. Cross-compiler (when you will have time - ie. ~ 2017... :-) )
Generally speaking about multimedia, isn't a high interest area for Delphi, imho, but
business graphics processing is. See all the buzz which is now around of Business
Intelligence and new forms of data display. For a neat demo and for what Delphi
SHOULD/MUST HAVE, (thanks Caps Lock!) see: www.tableausoftware.com
and take a product tour. (For the record, it has Firebird as embedded database engine).
Best Project for Communications: Winner: phpBB
- Delphi.NET can do it? How easy?
- IntraWeb can do it? How easy?
- Delphi 4 PHP can do it? How easy?
- For me, this shows, once again, that IT industry (programs, programmers,
internet/networks, hardware aso.) becomes more an more a mass, _social_ phenomenon
rather than a technical, restricted area.
Best New Project: Winner: eMule
(???? - eMule = "New"??? - well, compared with TCP/IP..)
Honorable Mention: Launchy
- From Delphi POV: Can I have a remote communication framework?
We have now a n-tier database communication framework (which of course must be
improved) and OTOH Indy for (rather) low-level stuff. Also we have web services. Can
I establish, _with_ease_, in a natural way, a n-tier informations channel ? Ie. not
only a stream but to send 'informations' (objects, data-types) featuring
error-handling, threading aso. OTOH, if it is a communication stream, then I want to
write it as a stream ('duck programming') ie. something like
with TRemoteStream.Create(<host>, [<protocol>]) do
begin
//also perhaps is good to have a mixin here in order to allow
//a threaded variant of this (hint, hint) - see Wikipedia & Python for ex.
WriteString('foo');
WriteInteger(10);
Free; //the RemoteStream
end;
- About Launchy: Beside that it is a very neat utility (highly recommended), imho, it
show us that we are in the incremental search era (partial match anywhere). Such a
control (a DDComboBox derivative) will be a welcome addition on the tool palette and
to be used in the IDE not only on searching for components (as it is now) but also on
the structure window and in the object inspector, showing in the drop-down window
only the items which match a certain pattern.
Best User Support: Winner: Firebird
- Again. They have several mailing-lists, a support scheme quite similar with
CodeGear's newsgroups. What it makes the difference, imho, is the firebird-devel &
firebird-architect lists in which users can speak about their feature requests, how
they would see the _next_ version of Fb. Also, there the core developers put their
problems, their thoughts asking for comments, help & community validation, having as
backend the JIRA's bugtracker. Here we're relying only on QC but this isn't enough,
imho. Perhaps we have something to learn from them (it's just a personal opinion,
though). Also, on this theme, see the O'Reilly's Open Source Awards (link on the top
of the message). The FIRST is Karl Fogel - Best Community Builder. Follow that link.
Highly recommended. He has also a free book. I found many wise things there.
Best Technical Design: Winner: 7-Zip
- Wow! Simplicity again. And efficiency. A very simple UI and a very powerful
compression algorithm. Today the needs of a strong compression algorithm are
increasing, rather than decreasing. (ie. the amount of data grows faster than raw
hardware speed). Imho, perhaps is effective to have in Delphi an implementation of
LZMA (AFAIK, there are some Pascal implementations{*word*154} around) - the zlib unit
isn't at today's standards any more (neither as interface, neither as compression
ratio) - even if it can prove useful. But, imho, the "rather than nothing" scheme on
a state-of-the-art product like (as it supposed to be) Delphi doesn't applies...
Most Collaborative Project: (The project most likely to accept your patches and value
your input) Winner: Azureus
- Yep. Anyone who goes at sourceforge.net/ will see that constantly Azureus is
on top 5 on "most active" (here in fact is most of the time 1st) _and_ "most
downloaded" projects. When someone listen community, the community listen back. Imho,
this is the leading factor in CodeGear's (re)rising. But it can be done better.
On the technical side, again a communications program. A RAD tool to build
communication applications based on Bittorrent protocol and/or on custom implemented
p2p (or C/S) protocols will make the day, imho.
Best Tool or Utility for SysAdmins: Winner: phpMyAdmin
- My lesson from here: OPF. Period. The 'Best tool for SysAdmins' and, even more, the
Sf.net comment on this: 'The project that keeps your network up while you sleep' is
clearly intended on a tool which manages a _network_ not a _database_ . But because
the winner is a _database_ tool - phpMyAdmin (which was also the 2006's winner)
confirms once more that the big problem nowadays isn't the network administration as
is the disruption between data from different sources (SQL, XML, text aso.) and code
so a global approach on a centralized management (and hence synchronization) between
data and code is quite important nowadays, imho. Hence Linq, which in a way, is
something which FoxPro/Clipper had in bare forms from years. OTOH, I saw these days
that one can do with DBX4 custom command types (see
blogs.codegear.com/PawelGlowacki/archive/2007/07/27/37764.aspx
and
blogs.codegear.com/steveshaughnessy/archive/2007/02/16/31865.aspx
)
which if it can be used with ease (I haven't yet time to explore) will be a very
powerful thing in reading other data structures a la Linq. Perhaps a CodeGear demo on
how we can read with ease, let's say, a tab delimited TStringList, having also data
types on 'fields' will be a light shower.
Best Project: Winner: 7-Zip
- Imho, this sums up the IT momentum in which we are. The users don't want anymore to
sit and learn. The want to use it right out-of-the-box. Nowadays, things must be
natural in order to be used. In a few amount of time they must learn a feature
otherwise they will throw it. That feature will cease to exist for them. And here, a
Pascal-like IT ecosystem, properly designed with language elements, data structures
and metaphors from real life (can) have a big advantage over a sscanf using {}, ^,
++,>>, || and lpszCryptlang (dot) nets we use today. Only to learn to (commonly)
leverage this.
Just my 2c,
(and hth)
--
m. th.