[Who says Apple is losing WebObjects folks at a high rate? What about all the
sights they are gaining? Is their rate of lossage any faster than any of the
others? Go review the threads regarding this-- there is no evidence that Apple
is *losing* sites at a faster/slower rate than anyone else. The Internet
industry is rife with turnover mostly due to the mismanagement of site
production that makes everyone unhappy-- it is easier to blame tools than
people and easier to change tools than people.]
What about:
- support?
- real database access? Does that include transaction management? editing
context like functionality? abstraction off the database layer?
- documentation? Can you hand it off to joe-blow-developer and have them
do anything useful?
- maintenance/installation?
- decent APIs? Or is it going to change/break with every release?
- flexible/extensible?
- abstraction between UI, business logic, etc...?
- ability to be used in a highly disciplined project process? I.e. am I
going to get screwed because the APIs don't work consistently and the
inconsistency can't be quantified?
---
As a concrete example, I can go to www.python.org and download a wonderful,
fully OO based, solution building applications that use the web as the user
interface. Why don't I?
It is an immature tool and does not have the refined elegance of WebObjects
that allows my company to build products using a highly refined process that
allows us to deliver a working solution on time. As well, there is no EOF--
nothing that encapsulates 10 years of indredibly refined evolution regarding
object abstraction on top of relational databases.
Again, the *cost of the actual license to develop and deploy using WebObjects*
is a trivial part of the overall cost of *any* application-on-the-web project.
Of course, if you aren't building a web based application-- or a site that has
the level of dynamic content that would require a solution like WebObjects
provides-- *don't use WebObjects*.
For a three month project involving four people-- 1 project manager, 2.5
developers and .5 quality/assurance resources at an average salary of $60,000
per annum, that breaks down into:
60,000 / 3 = 20,000 per person in salary
20,000 * 4 = 80,000 in salary alone for three months
Add: $10k to $100k in hardware costs-- we'll say $10k to be devil's
advocate.... another $25k for adminnistrivia (running a company is
expensive).... another $10k for miscellaneous-- software licenses, dev boxes,
etc (you are leasing these, aren't you?) and the total comes to:
$135,000
What does the extra $25,000 license for WO buy us?
Sure-- it bumps the cost up significantly. However, how much $$$ does it
save???
Well, what if using some garage tool that lacks documentation and support
pushes the development off an extra month? This is quite likely and happens
all the time with projects where it is difficult/impossible to quantify the
unknown. There goes an extra $20,000 in staff costs alone-- combine that with
the likelyhood that the staff has been busting ass to try and meet the
deadline, is stressed, and likely on the verge of burnout and you run the risk
turnover.
Turnover is expensive-- it'll take at least a week or two to get the new person
up to speed-- probably longer if you are using a garage style solution.
Worse, this will seriously cut into the productivity of the other developers--
they will have to take time to get the new guy/gal up to speed and share any
wisdom gleaned from the process. Again, using the immature or unrefined
solution, the amount of time sucked from the other developers will be much
greater because their will be more "common wizardry knowlege" that has to be
transferred by word of mouth-- you can't just give the developer a pointer to a
piece of documentation.
So-- considering the likelyhood of turnover, you are likely out another $5k.
Then there is the client or target audience.... you just missed a deadline.
Oops. What is that going to cost? Could cost a lot-- it'll likely mean that
you are now in an adversarial relationship with the client. That's bad.
Worse, when the project is delivered the client will probably shop around for
version 2 (or even version 1.1). If the client really feels screwed, they may
sue-- there goes another $5,000 to $25,000 in legal fees.
Ouch... the costs have really added up-- tangible costs are higher than the
initial license! Worse-- the intangibles can kill the company....
b.bum
Jeroen Clarysse wrote:
> I just phoned with a low-level programmer who has written his own webserver
> and stuff. He says that, for $25K, he can write a multithreaded, ultrafast
> UNIX C-app that does all the basic stuff for form-processing, database
> access and page-generation with a nice GUI. (using PowePlant, GUI is easy)
>
> in fact, he tells me that, for this price, any small programmers team can
> write a package that does everything a *lightweight* WebObjects would do.
> With no tpm-limit. "perfect for your small intranet solutions" as he puts
> it : "Handles 100hits per minute like a piece of cake".
>
> that summs it up pretty much I think why Apple is losing WebObjects at a
> high rate.
