How to identify a website's scale?

E

Erwin Moller

How do you know that it is a small project, a large project, or a
medium-sized project?
 
A

Arne Vajhøj

Erwin said:
How do you know that it is a small project, a large project, or a
medium-sized project?

I think that scale depends a lot on the company.

I would say <2500, 2500-10000 and >10000 hours but I suspect
Microsoft would add a couple of zeroes to each.

Arne
 
E

Erwin Moller

I think that scale depends a lot on the company.

I would say <2500, 2500-10000 and >10000 hours but I suspect
Microsoft would add a couple of zeroes to each.

Arne

Can you tell the size/scale after you understand a website's
requirements?
 
A

Arne Vajhøj

Erwin said:
Can you tell the size/scale after you understand a website's
requirements?

After reading requirements a good solution architect can
give an estimate in the +/- 30% range (assuming he knows
the domain).

After doing some high level design and analysis of
requirements a team can provide a +/- 10% estimate.

After the project is done you ask the accountants
for the number of milliseconds spend.

Arne
 
S

Stefan Ram

Erwin Moller said:
How do you know that it is a small project,
a large project, or a medium-sized project?

The adjective »small« is a natural language word
and intentionally vague. Therefore, any additional
precise definition will not catch its meaning and
use, which includes this vagueness.

In »small project« it has the meaning of »small« in English.
There is no specific Java meaning of »small«.
 
A

Arved Sandstrom

Erwin said:
How do you know that it is a small project, a large project, or a
medium-sized project?

Let's assume you've gone through an estimation process, and have some
numbers. Assume also that you have some reasonable formula (perhaps based
on history) for translating these numbers to effort...say, (# developers) X
(# months), so 3x3 or 5x6 or whatever.

It's in the context of your development organisation that 1x4 or 3x3 or 5x6
can be interpreted as "small", "medium" or "large". Even the accountants
can understand 3x3 when they know they have 12 coders total...that project
uses 25% of your people for 3 months, so it might count as "medium". A 1x4
might be "small", and a 8x12 is most definitely "large".

In no other organisation (except comparable ones) will these projects
necessarily be classified the same way. In a one-man shop a 1x12 is most
definitely a "large" project.

I find it better to use a (#D) X (#M) estimate for effort than total
man-months. You lose information with the latter. IOW, the same project can
use different total man-months depending on how many coders are assigned to
it.

AHS
 
A

Andy Dingley

How do you know that it is a small project, a large project, or a
medium-sized project?

Count the Lines of Code, silly!

Then divide it by Programmer Productivity, and you'll know how many
Man Months it will take to get it all working.

Add maybe 10% for contingency and you're done.
 
A

Andy Dingley

How do you know that it is a small project, a large project, or a
medium-sized project?

It's about the size and complexity of the codebase, not the volume of
traffic on it. Needing to cope with heavy traffic will affect you (an
ability to support server load balancing might make the notion of
"session" more difficult), but the basic volume of code is related
more to the features than to the number of users pulling repeated
pages. Some CMS sites can be huge, yet surprisingly simple internally
- it's usual that their editing features took far more work to develop
than their page-serving.

As to metrics for effort required, the only one that really works is
good requirements capture at the level of user stories (any less is
too vague, any more doesn't get done in time). Then do some level of
function-point counting exercise to estimate complexity, then apply
your last-project metrics for the same team, same tools and comparable
scale. Anything less than this gets progressively more wooly, in the
usual manner of old software projects.

Alternatively just call it a "small", "medium" or "large" project
arbitrarily and manage it on that basis. You'll get the results you
expected (keep focussed on what the most important rsult is, whether
that is time, cost, quality or features), maybe not the feature set
you hoped for. If you use an Agile approach though, this isn't too
bad, you're just not quite finished yet.
 
R

Roedy Green

is inversely proportional to the fourth power
of the time allotted for development.

Given that, you'd think large corporations would simply not put up
with "violin-making" style low level languages where every line is
lovingly crafted character by character.

You'd think the would demand the same brutal mass production applied
to goods.
 
R

RedGrittyBrick

Roedy said:
Given that, you'd think large corporations would simply not put up
with "violin-making" style low level languages where every line is
lovingly crafted character by character.

You'd think the would demand the same brutal mass production applied
to goods.

Don't they?

I expect that no software developers are involved in the manufacturing
of shrink-wrapped boxes of MS Word. I bet that all happens in some
mass-production factory in China.

I'd bet most software development is more analogous to something between
bespoke violin manufacturing and whatever the designers and
prototype-building craftsmen do at IKEA's design offices - not at all
like whatever the people/robots do in IKEA factories.

Just my äººæ°‘å¸ 0.02 worth.
 
D

Daniel Pitts

Lew said:
That pretty well sums up the Agile approach. It even applies to how
well one manages the Agile process itself: you're not doing too badly,
you're just not quite finished yet.

Corollary that I heard at a recent Agile Project Leaders' Network
meeting: "If you're still doing Agile the same way after a few weeks,
you're not doing Agile."
I would extend that. If all your teams are doing the same kind Agile,
you're not doing Agile.

If you have two distinct teams working on projects that are enough alike
to justify the *exact* same project management, then they are the same
project and the teams should either merge or work on something else :)
That seems to be the practice at my work at least.
 
A

Andy Dingley

Nothing like wrestling with doing it yourself to make you appreciate (or
deprecate) frameworks. (Apache Torque
<http://db.apache.org/torque/>
threw me off ORM layers so badly I almost didn't bother to learn Hibernate and
OpenJPA.)

Torque makes me wonder about changing career to become a watchmaker.
 
R

Roedy Green

How do you know that it is a small project, a large project, or a
medium-sized project?

When I was in grade 5, our teacher Miss Chapman asked me if a
hippopotamus was "big". I replied "compared to what? It is big
compared to a dog but small compared to an elephant."

I never heard the end it from my classmates about my stupidity.

The question might be more defined as, "in the class of land animal
species living or extinct is the hippopotamus one standard deviation
above the norm in volume."
 
R

Robert M. Gary

How do you know that it is a small project, a large project, or a
medium-sized project?

That's the difference between an experienced architect and a new guy.
Estimating projects and giving presentations are really what seperate
the men from the boys in every company I've worked for.

-Robert
 
A

Andy Dingley

When I was in grade 5, our teacher Miss Chapman asked me if a
hippopotamus was "big". I replied "compared to what? It is big
compared to a dog but small compared to an elephant."

Hippopotamus size conjugates as an irregular comparison, based on the
square root of the distance between the two of you, divided by the top
speed of any vehicle that you're in. If you're within sight, in any
sort of small boat, then they're big scary f*ckers, especially the
little ones with the pushy parents.

For an animal whose usual reaction to anything startling is to fling
dung at it with its tail,
the only surprise is that they haven't yet acquired Usenet access.
 
P

Philipp

Lew said:
I don't follow your logic. One would expect, if that formula held over
the entire domain, that companies would extend projects to the longest
feasible schedule.

Did he mean the total cost of the project or a cost per time?

If he meant total cost, this quote really makes no sense to me.

Phil
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Members online

No members online now.

Forum statistics

Threads
474,432
Messages
2,571,682
Members
48,796
Latest member
Greg L.

Latest Threads

Top