Embeded PHP/WEB/SQL program, was Re: So little actual software development in software

[markw-FJ05HQ0HCKaWd6l5hS35sQ at public.gmane.org (Mark Woodward)]

On 01/10/2011 12:00 PM, discuss-request-mNDKBlG2WHs at public.gmane.org wrote:
> Date: Mon, 10 Jan 2011 07:59:44 -0500 From: Matt Shields 
> <matt-urrlRJtNKRMsHrnhXWJB8w at public.gmane.org> Subject: Re: So little actual software 
> development in software engineering roles To: Stephen Adler 
> <adler-wRvlPVLobi1/31tCrMuHxg at public.gmane.org> Cc: discuss-mNDKBlG2WHs at public.gmane.org Message-ID: 
> <AANLkTikRfMgnTz3or_c112KV5s7GxM_73y5w1zRY3ii8-JsoAwUIsXosN+BqQ9rBEUg at public.gmane.org> 
> Content-Type: text/plain; charset=ISO-8859-1 On Mon, Jan 10, 2011 at 
> 6:30 AM, Stephen Adler <adler-wRvlPVLobi1/31tCrMuHxg at public.gmane.org>wrote:
>> >  For those of us who have lots of grey in our hair, this is an unsettling
>> >  situation. I've always wondered if such a group could come together and
>> >  work on a project for profit. Cell phone apps, stock trading software,
>> >  social networking plugins, what ever.. Or maybe one could form a
>> >  "boutique" software development group. I once worked for a start up in
>> >  Boston making a medical device. A software company in Canada was going
>> >  to charge the company 1.5 million for developing some imaging software
>> >  which I didn't think should cost 1.5 million, maybe several $100K.
>> >  Furthermore, the Boston start up charged a Japanese company 1 million +
>> >  for some software development which would have taking maybe several
>> >  months of two to 3 guys who know how to code complete. Is it just a
>> >  question of organizing such a group and advertising the groups
>> >  services?....
>> >
>> >  On 01/09/2011 11:01 PM, Mark Woodward wrote:
>>> >  >  I have been looking around for positions having had my last project
>>> >  >  canceled. I'm so tired, it seems like "software development" is more
>>> >  >  "software integration" these days. Maybe I'm old and washed up. I don't
>>> >  >  know, but jeez, I LOVE writing software. I mean, I love it. Problems
>>> >  >  wake me up in the middle of the night with rapturous solutions. I mean,
>>> >  >  seriously, I don't care what kind of software I'm writing, just be
>>> >  >  something that does something.
>>> >  >
>>> >  >  In the past few years, I've kept in touch with various high profile
>>> >  >  colleagues, hoping against hope that something interesting would show
>>> >  >  up. Sadly, no.
>>> >  >
>>> >  >  Venture Capitalists should no longer have the initials VC, it should be
>>> >  >  more like "vC." I know times are tough, but the whole VC deal is
>>> >  >  supposed to help you develop a product. The new deal is that you more or
>>> >  >  less have to have it developed. Worse than that, it needs to be fully
>>> >  >  buzzword compliant. While this is not an entirely new thing, it has
>>> >  >  become much worse.
>>> >  >
>>> >  >  Then, don't get me started on cloud computing. I mean, really, "cloud"
>>> >  >  computing. Talk about a buzzword. OMG then there's SaaS! None of these
>>> >  >  things are rocket science, and in many ways, they offer really powerful
>>> >  >  solutions to previously difficult or expensive problems, but not
>>> >  >  everything NEEDS to be cloud based. SaaS is a billing model, not an
>>> >  >  architecture, just ask skype.
>>> >  >
>>> >  >  Lastly, for various reasons, I'm a generalist. That means I have a
>>> >  >  pretty wide exposure with some really deep experience in a few areas
>>> >  >  like low level C/C++ and OS stuff (Windows, Linux, BSD). I could write a
>>> >  >  book on the various programming issues: threads, processes,
>>> >  >  synchronization, memory management, I/O, DMA, compilers, algorithms,
>>> >  >  pseudo-AI expert systems, databases, SQL variations, optimizations, and
>>> >  >  on and on.
>>> >  >
>>> >  >  What's the point?
>>> >  >  _______________________________________________
>>> >  >  Discuss mailing list
>>> >  >  Discuss-mNDKBlG2WHs at public.gmane.org
>>> >  >  http://lists.blu.org/mailman/listinfo/discuss
>>> >  >
>> >
>> >  _______________________________________________
>> >  Discuss mailing list
>> >  Discuss-mNDKBlG2WHs at public.gmane.org
>> >  http://lists.blu.org/mailman/listinfo/discuss
>> >
> The BostonPHP meetup group proposed a idea which might be good for people in
> this group too.  They talked about anyone in the group coming to a meeting
> with an idea or two.  The large group might split into a few smaller groups
> to develop a few prototypes, then out of the prototypes maybe develop full
> blown apps that they could make into commercial products.
>
That sounds like fun, I actually have something that's kind of cool in 
the works. While I'm looking, I'm using my spare time to work on a sort 
of backup project. I asked the list a while ago for some "must have" 
features. In the course of developing the system, I have a small 
sub-project that I may want to open source.

Here's the problem: I want to write the OS level stuff in "tight 
portable C/C++," but I don't want to do the UI development at the OS 
level, the UI should be user extensible and in a popular language, yet 
it has to be small enough to work on a small system (One of my target 
systems is an ARM based NAS device.) What do I do?

Using mongoose as the core web server, sqlite3 with encryption as the 
database, using a couple open source libraries as glue, I have a small 
embedded web server project that has a fully extensible architecture 
(using .SO files on unix and .DLL files on Windows) where you can run 
native modules as DLLs or php modules through CGI. (You could actually 
use any CGI system I suppose) The native modules are dynamically loaded 
at startup. The database interface is managed by a dynamic module using 
an HTTP GET/XML response protocol. The XML is formatted with hints that 
allow my PHP plugin (xmldbx) to read the response and produce a native 
PHP object.

For instance:
                 function Query($query)
                 {
                         $sql = urlencode($query);
                         $url = 
"http://$this->host:$this->port/sql?database=$this->dbname&query=$sql";
                         $file = fopen($url, "r");
                         $data = fread($file, 0x7FFFF);
                         fclose($file);
                         $obj = xmldbx_deserialize($data);
                         return $obj['rowset'];
          }


It currently works on Linux easily and the core libraries have compiled 
on Windows in the past. I will be getting this to work on next on 
Windows followed by x86 Mac.

Would the BostonPHP group have any interest?