struts-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Andrew Hill" <>
Subject RE: Struts Productivity Survey
Date Mon, 13 Jan 2003 04:09:06 GMT
30 people for 100 screen UI alone? What is this - NASA shuttle control via
web interface? ;-)

Our app (struts frontend, ejb backend) has 1.5 people doing a similar number
of screens (if you include listviews which thanks to the wonders of copy &
paste and inheritance take a good 2 minutes to implement...).

I  was given 3 days from a 'cold start' (ie: no jsp, struts, tomcat, j2ee,
ant - knowledge (luckily I was strong on html having done a fair bit of
domino web app stuff in a previous life)) to get the first business object
editing screen up. Ended up taking me a week (110 hours), as I needed a
couple of days just to figure out how to install everything and get across
some of the concepts (ie: servlets, request, session, struts actions, how to
keep state, etc...). Didnt use JSPs as I wanted something a bit more dynamic
for my view (started with XMLC but switched to creating my own XHTML DOM
based rendering code after the first iteration). (btw: didnt get that other
0.5 programmers till much later and still have to share him with another

The thing I found hardest to grapple with was the stateless nature of web
app design. I had been doing Swing UI stuff before this project, but now I
couldnt just instantiate a mob of widgets and add a few listeners (maybe I
should have used Barracuda? ;->). The stateless request & response stuff as
used in struts (and most web apps in general) is somewhat different...

The time after having everything installed correctly and when I was finally
doing coding (and a fair bit of study!!) to get that first screen up would
have been 2 or 3 days - though it wasnt exactly an example of good struts
coding! Next couple of screens similar, then after that it got a lot easier
as I started to actually understand what the heck I was doing! Nowadays it
takes me a couple of hours to do a screen (unless it has wierd
requirements - which they all seem to now!) most of which is nobrain typing.

I reckon I wasnt very productive the first 2 or 3 weeks, but after that got
more up to speed - especially once I had some kind of framework up to handle
some of the wierder common requirements specific to our app (and of course I
made a few minor mistakes I havent been able to factor out since :-<).

Nowadays only a small proportion of my time is spent doing basic struts
stuff for the screens - which pretty much comes automatically once you
understand the basics of struts anyway - the time consuming stuff is always
the wierd requirements and funny widgets that have to be implemented - and
that no framework helps much with as they are application & screen specific
(a good knowledge of html (& javascript if your permitted it) comes in very
useful here) - or the strange and twisted things one sometimes needs to do
to data before passing it to and from the B-Tier after input or output.

btw: can VB programmers really be trained to program, or is it true what
they say about the only thing being worse for coding than a VB developer is
2 VB developers??? (just kidding ;->)

-----Original Message-----
From: []
Sent: Saturday, 11 January 2003 01:49
Subject: RE: Struts Productivity Survey

Well, the question was somewhat, but not entirely hypothetical.  The 50 is
probably more like 30.  The problem is really another "my framework is
better than yours" debate and we're pushing to use a Struts-based framework
going forward instead of a home-grown.  One of my tasks is to quantify how
long it will take developers to get up to speed on Struts to estimate the
cost of moving over.

I learned Java, JSP/Servlets, and Struts pretty much all at the same time,
and it clicked easily for me.  But I learn new things easily and am not
intimidated by new things.  Others seem to have had varying levels of
difficulty picking it up.  The team that I worked on back then had little
trouble with it.

> -----Original Message-----
> From: Haseltine, Celeste []
> Sent: Friday, January 10, 2003 10:04 AM
> To: 'Struts Users Mailing List'
> Subject: RE: Struts Productivity Survey
> Greg,
> Wow, I don't envy your assignment.  There is one thing that I
> have learned
> over the years as both a developer and project manager,
> people learn at
> different rates, and pick up new skills very differently.
> In the case of the people who only know C++, the transition
> to Java will not
> be difficult.  If they have had no exposure to HTML and web
> development,
> then the transition to JSP's will take them a little longer.
> I've know some
> really good low level C++ programmers who just cannot get the
> hang of HTML,
> much less the tag library underlying struts.  These guys are
> better suited
> for the server side and EJB side of development.
> As for the VB programmer, your job is going to much more
> difficult.  Many of
> the VB programmers I have meet over the past 2 years have had
> no exposure or
> training in Object Oriented Programming (OO).  Many have
> moved to VB from
> mainframe jobs that they started back in the 70's and 80's.
> For those on
> your staff who have had exposure to C++ in college, the move
> to Java will
> probably be easier than for those on your staff you have come from a
> mainframe background.
> As to Struts, JSP's and the tag libraries.  You are looking
> at training
> these people in ALL three of these concepts/technology.  I
> would strongly
> suggest that you start out your training class in JSP's using
> the Model 1
> concept the first few days, and then introduce Struts and the JSP tag
> libraries for the duration of the training course.
> In all, my experience in training people in a combination of
> on the job and
> classes has been as follows:
> For those with heavy C++ experience, moving to
> Java/JSP/Servlets/HMTL took
> about month before these people were productive.
> For those with some exposure to C++ and OO, but no work
> experience, it took
> about 2 months for these people to be productive.
> For those with no exposure to C++ and OO (mainframe
> background), it took
> anywhere from 4 months to 9 months before these people were
> really useful
> and productive in the work environment.  A lot depended on
> the attitude of
> the person, and their willingness to learn new programming
> skills AND new
> programming/software concepts, particularly OO.  If you can
> get over the
> hurdle of the OO concepts, then you are 2/3 of the way there.
> If you add struts and HTML to the equation, I would add an
> extra 2 to 4
> weeks, before the staff is really productive and useful in the work
> environment.
> Keep in mind again that people learn at different rates, and
> that they learn
> in different ways.  The approach you take to exposing and
> training these
> people may not work for the entire group, and you may need to
> take extra
> time and try different approaches with some in the group.
> Good luck, this is one assignment that I don't envy you on.
> Celeste
> -----Original Message-----
> From: []
> Sent: Friday, January 10, 2003 9:40 AM
> To:
> Subject: RE: Struts Productivity Survey
> Ok, so let me pose it a different way.  Put yourself in this position:
> You're a consultant or an architect who has been tasked with
> building a
> Struts application with 100 screens.  You're actually migrating an
> application which was originally written as a fat client app
> in VB and the
> server was written in C++.
> You have about 50 developers who know VB and C++ pretty well.
>  Some of them
> know Java/JSP/Servlets very well.  Some of them know it very
> little.  None
> of them know Struts.
> Your job is to estimate the cost of getting these folks up to speed on
> Struts.  You already have the application design.  THere's
> going to be a
> thin business layer called by the Action classes that will
> use EJB session
> beans to provide application functionality.  Fairly simple.
> You have to
> train these people to use Struts, create actions and pages,
> and write the
> web-tier business layer.  Somebody else is doing the EJB stuff.
> How do you estimate that cost?
> How long do you think it will take for these various
> developers to write
> their first business function?
> How long to do the next one, etc?
> > -----Original Message-----
> > From: Andreas Mack []
> > Sent: Friday, January 10, 2003 12:34 AM
> > To: Struts Users Mailing List
> > Subject: Re: Struts Productivity Survey
> >
> >
> > On Fri, 2003-01-10 at 07:57, Rick Reumann wrote:
> >
> > > easier. I think it's really going be difficult to get an
> > accurate feel
> > > for how long it takes people to get 'up to speed' with
> > struts since I
> > > think it's much easier now for new developers to learn
> > struts simply due
> > > to the books and better documentation available.
> >
> > I agree 100%. I looked at Struts for the first time in April
> > 2001 mainly
> > for the forms stuff, really wanting to use it. I've read
> the UserGuide
> > and said "What is he talking about!?" Half a year later, with a real
> > project at hand it went much faster, using the /example stuff. Back
> > then there were no DynaForms, no Tiles, no Nested, all the
> stuff that
> > makes things much easier now. The pages that are now the Taglib API
> > Reference were the best resources back then.
> >
> > Greets,
> > Andreas.
> >
> > --
> > Andreas Mack <>
> > mediales. GmbH
> >
> >
> > --
> > To unsubscribe, e-mail:
> > <>
> > For additional commands, e-mail:
> > <>
> >
> >
> --
> To unsubscribe, e-mail:
> <>
> For additional commands, e-mail:
> <>
> --
> To unsubscribe, e-mail:
> <>
> For additional commands, e-mail:
> <>

To unsubscribe, e-mail:
For additional commands, e-mail:

To unsubscribe, e-mail:   <>
For additional commands, e-mail: <>

View raw message