[OT] mod_perl longevity [Was: mod_perl advocacy project resurrection]

[Ajit Deshpande <aj...@skycorp.net>]

On Tue, Dec 05, 2000 at 11:34:49AM -0800, brian moseley wrote:
> i had lunch with doug and jon swartz not too long ago,
> talking about the possibility of starting a web application
> infrastructure company based on mod_perl and mason. when we
> got down to it, the fundamental question was: why not just
> use java? and we couldn't find any answer other than "i like
  ^^^ ^^^^
> perl better". and that's not a reasonable business
> justification.

Well, the above question pre-supposes that Java is inherently
*better* than mod_perl for some definition of "better". 

Without getting into a long-drawn holy-war, I dont like the idea 
of "one size fits all" when it comes to web application 
infrastructure technologies. There are and will be perfectly good
reasons to use mod_perl / Java under the reasonable assumption that 
both technologies will evolve in the foreseeable future.

Ajit

Re: [OT] mod_perl longevity [Was: mod_perl advocacy project resurrection]

[Ajit Deshpande <aj...@skycorp.net>]

On Tue, Dec 05, 2000 at 12:28:08PM -0800, brian moseley wrote:
> On Tue, 5 Dec 2000, Ajit Deshpande wrote:
> 
> > Well, the above question pre-supposes that Java is
> > inherently *better* than mod_perl for some definition of
> > "better".
> 
> it's true. i stayed away from defining better in that msg,
> but explored in a separate one in this thread. suffice to
> say, the wealth of existing standards-based components, the
> focus on creating a platform for enterprise applications,
> and the competitive vendor landscape for development tools
> and infrastructure components, all of help define "better"
> imo.

well, there is always going to be a niche market for cleverly 
hand-crafted web-infrastructure solutions. personally, thats
the space i am interested in.

also, imo, the notion of building enterprise-ready app-infrastructures 
overnight by deploying infrastrucutre components from some big-name 
vendor may not always be feasible.

the size of the pond where people can get by with off-the-shelf
components is defintely bigger, i would agree. but you are never 
gonna find the Amazons, the Yahoos, the Hotmails and the Valueclicks 
get by with those. 

so, if you, doug & jon start a company, dont you think every new 
mod_perl based nich web company is gonna call on your combined 
expertise! *dammit* start the company already! (and then hire me 
to do the grunt mod_perl work :)

ajit

Re: [OT] mod_perl longevity [Was: mod_perl advocacy project resurrection]

[brian moseley <bc...@maz.org>]

On Tue, 5 Dec 2000, Ajit Deshpande wrote:

> Well, the above question pre-supposes that Java is
> inherently *better* than mod_perl for some definition of
> "better".

it's true. i stayed away from defining better in that msg,
but explored in a separate one in this thread. suffice to
say, the wealth of existing standards-based components, the
focus on creating a platform for enterprise applications,
and the competitive vendor landscape for development tools
and infrastructure components, all of help define "better"
imo.

[OT] Starting a Company on OS [Was: mod_perl longevity] [Was: mod_perl advocacy project resurrection]

[Gunther Birznieks <gu...@extropia.com>]

I thought I would change the topic yet again. :)

At 02:58 PM 12/5/00 -0500, Ajit Deshpande wrote:
>On Tue, Dec 05, 2000 at 11:34:49AM -0800, brian moseley wrote:
> > i had lunch with doug and jon swartz not too long ago,
> > talking about the possibility of starting a web application
> > infrastructure company based on mod_perl and mason. when we
> > got down to it, the fundamental question was: why not just
> > use java? and we couldn't find any answer other than "i like
>   ^^^ ^^^^
> > perl better". and that's not a reasonable business
> > justification.

That's a very techie thing to think about. I'll stay out of this part of 
the conversation as I said my piece on it in several other threads over the 
past 2 years. :)

I would be very careful about setting up a business entirely around 
mod_perl for several reasons (couched in an anecdote about my experience 
starting a company).

We started a company around our open source products 8 months ago and left 
the cushy full-time job sector. While I don't regret starting the open 
source company I have say several things that we discovered (really common 
sense but)...

1) Even if you hire a good business team to back up you as techs, at least 
one of you will not be coding or engineering anymore. You will be business 
development/sales/marketing/providing support in writing and revising the 
business plan, making the next strategic alliance, closing the next big 
deal (big deals take a lot of time to close both length of time and 
building the relationship).

The mod_perl community will definitely lose one of you guys for sure.

2) Service/Consulting contracts extremely time consuming to administrate if 
you want to manage it "right".

3) It's a bad idea to make a company dependent on the fact that you have 3 
really hot-shit developers because

(A) one of you will be taken away for business things, almost guaranteed
(B) Your clients will want YOU to work on it not someone you hired and 
trained -- because until they've been working on mod_perl for several 
months, it is unlikely they would be doing work. In the meantime you pay 
someone a salary to learn enough to be good enough (so prepare for a 1-3 
month buffer of paying salary).
(C) Service based company's make money off of headcount not off of a few 
good people.

4) Unless you are injecting a lot of cash into your company or know someone 
who will, you will not reach a critical mass of directly billable headcount 
necessary to generate revenue which will take care of the "overhead" you 
need to hire to grow a company beyond a couple of people (CEO, Business 
Development, COO, Marketing, WebMaster/Graphics guy (invariables 
infrastructure contracts ask if you can also take on a tiny bit of apps 
work and then it will also be an issue of doing graphics for their site) 
and CF0) maybe in that order ... but that's another room for debate that's 
not for this list.

It's really hard to do this on your own money. We started our company with 
5 people. 2 techs, 3 business. We've grown in 8 months to 15 people. But 
the process has not been easy. The first 6 months definitely took me away 
from posting here (some would say that's a good thing) and it's completely 
taken away the other tech who I started with as he is still mired into 
doing business stuff.

If he has spare time it isn't to do open source as much anymore as to code 
to generate revenue so we can hire more people.

Of course, getting funding is cool (the VC route), but we haven't done that 
route. The VCs we've gone to all want over 51% of your company to give you 
money (especially these days)... Losing control sucks when you are doing 
something you are passionate about (open source).

And they also want to see that your team has run a business from scratch 
before. This was a minor problem for us when we started looking at funding 
opportunities, although not anymore as the same VCs are now impressed we 
survived and prospered without funding at a time when dotcom's have been 
dropping like flies.

Dotcom Funerals are really hard to watch and demoralizing also. I would not 
want to see a mod_perl-based dotcom funeral because some VC was controlling 
the company and that VC doesn't understand open source.

Of course, this may be sour grapes, but at this stage, although funding 
would be fantastic, I shudder to think that if we had funding from the 
wrong VC or investor then we would be screwed.

Anyway, the summary of this story is that running a business if really 
hard. I thought going into it that quiting my full time job to work in an 
open source company would mean more time to do open source and advocacy.

Boy was I wrong. The process of starting a company took me out for 6 
months, and only now am I getting back into it. I love doing the company 
thing and I've learned a lot about real business (especially making the 
hard decisions that you have to when you dont have gobs of money), but it's 
still completely not what I expected as a techie.

Later,
    Gunther