RE: [Cxtable-devel] RE:Cygwin...etc..[continued]

[Borne Goodman-M. <bm...@eg...>]

I seem to be having some odd problems with your email not wrapping
correctly in my responses, but I will do my best to respond to all of
your questions / comments, and tell me if I miss anything.

It is indeed difficult, when dealing with a project that doesn't have
any customers to drive functionality, to determine what to add next, and
I have this same problem with my OMInit project.  The process I usually
use is to determine what functionality I think would do the most to
bring people to the product, and implement that first.  Once there is a
large user base, then you have people asking you to make changes, and
unless there is a re-factoring or other internal thing which take
priority, you can just go down the user list of wants.

From that perspective I think that a nice UI revamp would be of the
highest priority.  The most important part of any program to any user is
the UI, because that is all they understand of it.  If the UI doesn't
look good, or doesn't react well, they won't use it.

Before you run off and start on the new UI, I also want to mention that
any structure built on an unstable foundation is prone to fall faster
when stress is placed on it.  If there are any lingering communication /
architecture issues, I would suggest fixing these before anyone uses the
code.  Once the core is stable, start adding things onto it, and if
designed well, it shouldn't be difficult.

Most of the people I know (including myself) work from behind a firewall
at home and / or at work.  The ability to run the program in this
environment is VERY important.

I would then suggest the file sharing capability, as p2p file sharing
which works well is always a great way to grab a lot of users.

The XML / DTD change would be trivial, I believe, if done in the same
way I do it for OMinit, but older versions of the code would no longer
work with the new code, unless we just deprecate the old mechanisms, but
since the user base is so small, this wouldn't hurt at all right now. 
Because of this, the change should probably be done before a large user
base is established.

It is also very important, before a lot of people are using it, to
finalize the Plug-In support, and development of plug-ins is not
necessary for this.  As more and more users user the product, 3rd party
developers will use the plug-in support to make their own, if the API is
a nice one.  The Plug-In API is VERY important, and I can equate it with
the gaming environment on Windows.  The reason for a high percentage of
Windows systems is because that is the only environment you can play
most games in, and because most users run in that environment, game
developers write code for it.  This is a cycle, and one you want to
build for the product.  Write a great API for plug-ins and get a decent
user base, and developers will write programs for it, and if they are
good, users will user your program only because that plug-in is
available.  This is cyclical and the beginning of it requires only
plug-in support, and a decent user base.  I feel in this paragraph I at
least matched your rambling and have shown that I can do it myself :).

Due to the above statement, I think the need for additional plug-ins to
be written should be on the bottom of the list of things to do, even
though they sound like cool ones, and with the above finished, can be
accomplished next.

I am not easily scared, either by the size of a project, of the lack of
size of a project, as each idea, with work, will naturally become of a
proper size in respect to it's worth (given enough time).

I have found myself recently (when not at work) preoccupied with my own
OMInit project, and because of some of the abilities it has, and some of
the new JDK 1.4 APIs I want to play with, I have been forced to work on
that more rather than on this.  I find the 2 spaces very interesting,
that of a server that can handle a huge amount of traffic for a single
reason, and large number of peers which can also handle a huge amount of
traffic, but in a different way and for multiple purposes.  

I should, sooner rather than later, look more closely at the current
cxtable architecture to validate that it is a solid foundation to more
forward upon, try to pinpoint some of it's weaker spots, and perhaps
even strengthen / replace them myself.  I have a few more changes to
make to OMInit before the initial chatting capabilities are code
complete and I shall halt, for a time, on that project, so that I can
look more at this one.

Most of my work in the past has been in the same space as the OMInit
project, and that is handling lots of client's access to a set of
limited resources.  I like cxtable mostly because it is a similar
problem, and just a different approach at a solution.  I have rather
grandiose dreams myself, and would like to find that perfect medium
between peer to peer computing and client / server computing and need a
p2p environment to work in to advance that dream.

You are very welcome, and thank you,

--bjgm 

On Wed, 2001-12-19 at 15:34, Williams, David wrote:
> Woops..hit "Send" too early on that one...
> 
> I am at an interesting state in this development...  I have a bunch of little and big unfinished cxtable projects...  And I can't decide which to do..
> 
> Naturally...the CVSing of the code has to be complete...and the associated packaging issues...  But after that, I am not sure where I had wanted to go...(FIRST, most important outstanding cxtable project)
> 
> The only "plugins" that I currently have thought of are game related... and that is not the only exclusive capability of such a network....  However... I am needing to come to grips with non-global multi-"user" plugins..(A SECOND, importance arguable either way, outstanding cxtable project).
> 
> That is, a plugin that does not exist for every person from the get-go (non-global) yet would involve more than a simple peer-to-peer (such as Chess).
> 
> As far as that is concerned, two plugins come to mind.  A Poker game, and a new game called xTournoChess, which is a 16 X 16 board with 4 Chess players instead of two, one on each side of the board...
> 
> The idea, then, is implementing the plugin so that either one of two things happens.. -1- There is a plugin-server process running on a machine that mediates between the various clients involved, and runs the plugin.. Thereby requiring a new connection to that server-process from each of the peers involved....OR -2- The initiator of the plugin runs the plugin-manager piece, and all of the communication occurs through the initial p2p connections....
> 
> Does the above make sense?  I first thought of -1-, but gravitate heavily now toward -2-...  Because, given a set of people (called A,B,C,D,E)...  There are two connections between each, and these are readily able to accept messages... tagging the message as "<PLUG><plugname-manager>" instead of "<PLUG><plugname>" would allow the manager piece access to the instructions, which it could then propogate to the involved parties...
> 
> This would allow the plugin "logic" to be executed once, and the appropriate moves/instructions to be propogated to the 'screens' or other implementation of the plugin elsewhere...
> 
> 
> That is one among many of the cxtable projects outstanding... yes...it is game-related as far as implementation, but it would allow the infrastructure to be built for multiple-peer plugins that would extend beyond the "tournochess"...
> 
> Other cxtable projects (a THIRD, and equally arguable either way as far as importance is concerned, outstanding cxtable project) are the potential DTD-ing of the messaging, so that a robust parser can be used to process the XML that is passing between the peers/etc...  This is also a potentially important project...although cosmetically from the outside it would not reflect much difference (ideally ;-) ) it would greatly impact certain scalability issues in the future...
> 
> Another piece (a FOURTH) is a file-sharing (plugin?) piece... This could be run as a "download-server" on each peer, at a dynamically assigned port...  The ftp-client could be configured to read to that port, create a data sock, and download... I have a piece of software (currently being used by the Registry Server to publish itself) called Linlyn that I already modified as xDownloadLinlyn, with an interface xDownloadListener...  Multi-file downloads would create an instance of xDownloadLinlyn, which would call the xDownloadListener's method to notify it (filename,contents)... Either way...the concept of file-sharing to be plugged in is another work...
> 
> Yet another (FIFTH, and potentially of second importance only to CVS, depending on who you are) is the Swing conversion.  'nuff said.
> 
> Still another (SIXTH, and also potentally of second importance only to CVS, depending on who you are) is the refining of the GUI (swing or not) and the user processes so that they are more friendly, and more automated... 
> 
> Still yet another (SEVENTH, and also potentially of equal importance to CVS, swing and GUI-refine) is the setting up for behind configurable proxies/firewalls.  
> 
> 
> I hope I neither scare you off from the huge set of projects I embark upon, nor scare you off from the still potentially limited list I myself can come up with...
> 
> When I advertised for help, I didn't know what to expect... I had dreams for this project...and some of them have come to fruition...and I have learned a great deal as a result of simply working on the project...  You seem to wish for your role to be that of mentor, and that is perfectly acceptable to me.... ANY level of involvement is great... Code, help organize, dream, spark new ideas...whatever...  The reason (mentioned in the help-wanted ad but re-iterated here for clarity) I wanted another developer is more for vision and less for individual lines of code...  You certainly have exceeded all of my expectations as far as a team member is concerned...
> 
> One thing I might or might not have already demonstrated is my ability to ramble...
> 
> If I hadn't prior to this email, I'm sure I did with this verbose email..
> 
> Maybe I could ask you a semi-frank question...  What is that you saw in the project that made you want to work on it..?  ((There are no bad answers ;-))
> 
> Do you have any thoughts on the above projects?  Do you have any more to add that you think might be of more importance.. (All opinions weigh heavily here)
> 
> Thanks,
> Dave
> 
> David Scott Williams
> Computer Associates
> Marketing Representative-Sales Call Center
> One Computer Associates Plaza 
> Islandia, New York  11749
> tel: +1 800-243-9462 ext. 73431
> tel: +1 631-342-3431 (Direct)
> fax: +1 631-342-5734
> wi...@ca...
> 
> 
> _______________________________________________
> Cxtable-devel mailing list
> Cxt...@li...
> https://lists.sourceforge.net/lists/listinfo/cxtable-devel
>