You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@tapestry.apache.org by Bill Lear <ra...@zopyra.com> on 2003/03/18 14:20:23 UTC
Vlib code distribution
I remember somewhere someone remarking that Tapestry would tend
to produce applications with fewer lines of Java code.
Looking at Vlib, I see about 8,500 lines of Java code (I excluded
copyright lines), 1,000 lines of page and jwc code, and 2,500 lines of
html.
That's about 12,000 lines in all, of which, very roughly, about 70% is
Java code, 10% is page/jwc code, and 20% is html.
The Tapestry page asserts that "Graphic designers and Java developers
can work together", and assuming that page/jwc code is handled by Java
developers, that's about 80% of the work handled by Java developers,
and 20% by HTML developers.
So, for every HTML developer, I'll need 4 Java developers. This, of
course assumes roughly equal measures of "work" and productivity
across the two domains.
Does that seem about right?
Bill
RE: Vlib code distribution
Posted by "Howard M. Lewis Ship" <hl...@attbi.com>.
I'm in the embarrasing (?) position of now having done a large-scale
Tapestry project in quite some time. I left Primix for WebCT with the
expectation that we'd be hip deep in Tapestry by now, but that hasn't
materialized (at least, not yet). Mind Bridge, Adam Green, Eric Schnieder
and others on this list will probably be more helpful.
--
Howard M. Lewis Ship
Creator, Tapestry: Java Web Components
http://jakarta.apache.org/proposals/tapestry
> -----Original Message-----
> From: Bill Lear [mailto:rael@zopyra.com]
> Sent: Tuesday, March 18, 2003 9:06 AM
> To: Tapestry users
> Subject: RE: Vlib code distribution
>
>
> On Tuesday, March 18, 2003 at 08:28:37 (-0500) Howard M.
> Lewis Ship writes:
> >Your underlying assumption: that one line of HTML (by an HTML
> >developer) and one line of Java (by a Java developer) takes the same
> >amount of time to code and debug is completely flawed.
> >
> >My experience has been about 1:6 (one HTML developer works
> with 6 Java
> >developers), though typically not all of the Java developers are
> >working on the presentation layer. This matched the overall mix at
> >Primix at the time, which had about give HTML/graphic
> artists and about
> >30 Java developers.
>
> Interesting: if your estimate is about 6:1, that's not too
> far off of my estimate of 4:1. So, if your application at
> Primix had roughly the same mix of work for Java and HTML
> developers (80/20), you would have had 4 units of work
> handled by 6 Java developers, and 1 unit of work handled by
> an HTML developer. That's about 2/3 unit of work per Java
> developer, not too far from the 1 unit of work per HTML developer.
>
> So, if I have 10 units of work divided in the above
> proportions, I'll need 8/(2/3) Java workers and 2/1 HTML
> workers, or 12 Java versus 2, exactly your 6:1.
>
> Does this sound about right? The issue here is what the
> productivity measurement should be: is 2/3 to 1 about right?
>
> I'm looking to estimate and justify the needs of our project,
> so this is important, though of course, I'm just looking for
> back-of-the-envelope estimates ...
>
>
> Bill
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: tapestry-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: tapestry-user-help@jakarta.apache.org
>
RE: Vlib code distribution
Posted by Bill Lear <ra...@zopyra.com>.
On Tuesday, March 18, 2003 at 08:28:37 (-0500) Howard M. Lewis Ship writes:
>Your underlying assumption: that one line of HTML (by an HTML developer) and
>one line of Java (by a Java developer) takes the same amount of time to code
>and debug is completely flawed.
>
>My experience has been about 1:6 (one HTML developer works with 6 Java
>developers), though typically not all of the Java developers are working on
>the presentation layer. This matched the overall mix at Primix at the time,
>which had about give HTML/graphic artists and about 30 Java developers.
Interesting: if your estimate is about 6:1, that's not too far off of
my estimate of 4:1. So, if your application at Primix had roughly the
same mix of work for Java and HTML developers (80/20), you would have
had 4 units of work handled by 6 Java developers, and 1 unit of work
handled by an HTML developer. That's about 2/3 unit of work per Java
developer, not too far from the 1 unit of work per HTML developer.
So, if I have 10 units of work divided in the above proportions,
I'll need 8/(2/3) Java workers and 2/1 HTML workers, or 12
Java versus 2, exactly your 6:1.
Does this sound about right? The issue here is what the productivity
measurement should be: is 2/3 to 1 about right?
I'm looking to estimate and justify the needs of our project, so this
is important, though of course, I'm just looking for back-of-the-envelope
estimates ...
Bill
RE: Vlib code distribution
Posted by "Howard M. Lewis Ship" <hl...@attbi.com>.
Your underlying assumption: that one line of HTML (by an HTML developer) and
one line of Java (by a Java developer) takes the same amount of time to code
and debug is completely flawed.
My experience has been about 1:6 (one HTML developer works with 6 Java
developers), though typically not all of the Java developers are working on
the presentation layer. This matched the overall mix at Primix at the time,
which had about give HTML/graphic artists and about 30 Java developers.
I'd be interested to hear what some of the other groups out there using
Tapestry more recently can say. Some of the newer features (library
support) simplify things, and I feel that 2.4/3.0 is a whole new ballgame.
--
Howard M. Lewis Ship
Creator, Tapestry: Java Web Components
http://jakarta.apache.org/proposals/tapestry
> -----Original Message-----
> From: Bill Lear [mailto:rael@zopyra.com]
> Sent: Tuesday, March 18, 2003 8:20 AM
> To: tapestry-user@jakarta.apache.org
> Subject: Vlib code distribution
>
>
> I remember somewhere someone remarking that Tapestry would
> tend to produce applications with fewer lines of Java code.
>
> Looking at Vlib, I see about 8,500 lines of Java code (I
> excluded copyright lines), 1,000 lines of page and jwc code,
> and 2,500 lines of html.
>
> That's about 12,000 lines in all, of which, very roughly,
> about 70% is Java code, 10% is page/jwc code, and 20% is html.
>
> The Tapestry page asserts that "Graphic designers and Java
> developers can work together", and assuming that page/jwc
> code is handled by Java developers, that's about 80% of the
> work handled by Java developers, and 20% by HTML developers.
>
> So, for every HTML developer, I'll need 4 Java developers.
> This, of course assumes roughly equal measures of "work" and
> productivity across the two domains.
>
> Does that seem about right?
>
>