[LINK] CSIRO quizzed on $9.47m web site

[Craig Sanders]

On Thu, Feb 17, 2005 at 09:25:17PM +1100, Bernard Robertson-Dunn wrote:
> Craig Sanders wrote:
> > a million dollars would buy you a massive load-balanced web server
> > farm AND a year or two's worth of salaries for a sysadmin and asst,
> > 3 or 4 programmers, 3 or 4 web monkeys, and 1 or 2 people to liase
> > with the various CSIRO projects/depts/groups and help them get
> > their stuff onto the main site.  might even buy you a good technical
> > manager(*) for the project too - if not, under $100K or so would.
                                             ^^^^^
oops, typo. i meant "another"

> snip more technology stuff...

except for one of the postscripts at the end, NONE of what i wrote was
"technology stuff".

but, contrary to what you are implying, this nasty "technology stuff" is
important.  without it, your web site isn't going to work.  without
understanding it, you are going to get ripped off by almost any consultant you
hire.  technology is not something you can just ignore as if it's a trivial
detail.


> With the greatest respect, I would suggest that you don't understand why
> these projects cost what they do. 

actually, i do.

they cost so much because they get outsourced to useless parasites
who charge like wounded bulls.

> Secondly, in a project like this, the cost of the IT is only a very
> small component. 

yes, that was part of my point.

> A significant part of the cost is in development time of the various
> web sites. Consistency of design, look and feel all take a significant
> amount of time to produce.

yes, but that is grossly inflated when the job is outsourced to parasites with
their snouts in the trough.

outsourcing any IT project will inflate the cost by at least a factor of 10 -
partly because the bullshit economic irrationalist theory behind outsourcing
also involves sacking any in-house staff who might be capable of understanding
the technology enough to prevent being ripped off.

> Have you had a look at the site map?
> http://www.csiro.gov.au/index.asp?type=blank&id=site_map
> 
> There are 215 <a href=> statements. I followed one at random and it had
> 35 further links, one of those had 45 further links.

so?

> If that were typical, there would be over 300,000 pages on the CSIRO
> site.

so?

> If each page required one hour of human time for development, checking,
> testing and putting into production, (a number which I suggest is very
> low, even given a Content Management System) that's 3million person
> hours. At $50 an hour thats $150 million

really, all this data is available on other web sites - the original sites
that each dept or project had before they were all merged into one big site.

only an idiot (or a parasite consultant, but i repeat myself) would retype it
all.  anyone with any sense would write a script or set of scripts to import
it (and do whatever layout & format conversions are necessary).  BTW, this is
yet another advantage of free software - you can do this kind of thing because
the internal workings are documented and available, not treated as some kind
of proprietary trade secret which customers are never allowed to see in case
they "steal" the "intellectual property" (or realise how shoddy the product
actually is).

even in the unlikely event that it was impossible to import and just had to be
retyped, you don't pay $50/hour for that.  you pay shit-kickers $12-$15/hour
at most - and an organisation like CSIRO could rope in students to do it for
free.

or you create a design, define the standards and whatever templates you need,
run some training sessions, and get whoever did the original web sites in each
department to cut-and-paste and retype and re-layout the content from the old
site into the new CMS site.  they know their content, they know what's
important, they have access to the source material that the original site was
developed from, and they'll be motived to get it right.

and if your CMS can't cope with 250 people editing their sub-sites
simultaneously, then you've chosen the wrong hardware and software
combination.  start over with something that WILL do the job.

> That was the cost of web page production. What costs even more is the
> process of content development (ie what information should go on the
> page) That requires the involvement of subject matter specialists - and
> that costs even more.

that content already exists.

> $9.6 million strikes me as a bargain.

it strikes me as yet another instance of parasitic corporate welfare.

> The problem with letting IT people run IT projects is that they don't
> understand that the IT cost is a very small fraction of the total
> project cost. When managers (and politicians) believe the IT people the
> chances of a project being perceived as a success are really small.

the problem with letting consultants run IT projects is that they're only in
it for the money.  they don't give a damn about what they're doing, or how
they're doing it, or even whether it will actually work (if it doesn't work,
they just get paid more to "fix" it) - all they care about is how much they
can screw out of the project. 

> Having been involved in a project that redeveloped a Federal Government
> web site with a punny 20,000 pages, I would claim some small experience
> in this area.
> 
> What's yours?

probably a lot more than yours.

i spent the last 7 years as senior systems admin for a govt. owned ISP which
concentrated on hosting government and NGO/community-group sites, and
developed huge govt project sites.  we had some of the largest sites in
australia, and probably the most comprehensive portal site in aus covering
anything and everything about victoria and australia (and many other topics).

i know what it costs, and i know what it takes to develop large websites, and
how to design systems that will scale to take enormous load  - and i know it
from the perspective of having actually DONE it, not just contracted some
parasite consultant to do it and believing whatever bullshit they give me to
justify their outrageous overcharging.


you don't have to spend stupid amounts of money to get a good result.  in
fact, you're more likely to get a good result if you spend a sensible amount
of money - and spend most of it on good people, who can get the most out of
hardware and software.  i.e. don't spend millions on some crappy CMS package
so you have only peanuts left for staffing - instead hire good people who have
enough clue and experience to get the job done.


i've seen dozens of govt. depts and NGOs and community groups and so on get
ripped off blind by some parasite consultant who sees them as a gravy train.
i've seen it happen over and over again.  they're easy to rip off - they know
next to nothing about computers or the internet and they'll believe pretty
much anything they're told.  i have developed a serious loathing for the kind
of scumbag leech that rips off organisations who put their heart and soul into
helping their community rather than exploiting it.

craig

-- 
craig sanders <cas at taz.net.au>           (part time cyborg)