The other day I saw a posting in comp.groupware.lotus-notes.programmer, where the poster said that he have been defending Lotus Notes from people that dislike it and want to get rid of it. I asked what problems the users had, and this is the response:
This is the only application we have in Notes - we do use notes mail -
but it is the application that is giving Notes a bad name.
It is the result of a extremely bad design and implementation of an
application 12 years ago in Notes 4.5.
The application - a customer order configurator - was created by
someone else in a branch of the
company which sold us off so the designer/programmer is no longer
available to us..
This has been compounded by massive changes over the last 6
years. I have added 2 new databases (6 databases now) and possibly
140,000 lines of code to the application over the last 6 years. However
this has been hampered by the original bad design.
This results in users getting annoyed rather frequently.
So basically, one badly written application, probably developed by someone that was not a programmer or did not understand Lotus Notes, then patched over 12 years is giving Notes a bad name... What can be done about things like this? There are some (old) best practices out there, but perhaps it would be a good idea to write some new ones.
0
0
3
1 Ian W Randall Permalink At least this guy didn't blame the Lotus Notes application when his printer ran out of paper.
2 Nathan T Freeman Permalink 140000 lines of code!??!?!
That's about 22000 lines of code a year. The longest script library
I currently have active is about 2000 lines long -- so this guy
made the equivalent of 10 of those a year!?
This screams to me that there's a whole bunch of really, REALLY bad
code in there. Probably a lot of NotesDatabase.Search lines
followed by .GetNthDocument and some NotesDocument.~$Readers =
"yadda"
No custom application on a PC platform can survive 12 years without
a refactor and still be considered efficient. If you built a
website in 1996, would it's operation be state of the art today? If
you worked with a native OS at the time it would be Windows 95! The
biggest competing platforms for coder attention in the IT workplace
were Visual Basic and PowerBuilder, not .NET and J2EE.
It sounds to me like this is a company that has never considered
the idea of refactoring the app, whether on the Notes platform or
some other platform, for the good of their users. If it's an order
configurator that likely runs their business, they probably endure
tens of millions of dollars of lost productivity for users of the
app simply because they don't want to invest in it.
And that's the lever that this guy needs to use: "Here is our
opportunity cost by working with this essentially broken
application for the last decade. Here's what it would cost to
modernize it. As you can see, the ROI is about 2 months."
Hell, I'd go to the CFO with alternative platform pitches as well.
Even if you have to suggest jumping platforms, at least run the
numbers -- because even a switch to J2EE would probably have strong
ROI for them at this point.
I've known plenty of customer situations like this, and usually the
reason why there's no proper refactoring is because a) no one has
shown ROI impact; or b) the company believes it's working on
something that's replacing the aging system already. (I know one
customer that refuses to deploy substantial changes because the
Notes system is going away in 6 months -- they've been making that
claim for at least the last 5 years.)
Man, a basic design set out in 4.5. So Lotusscript was in its
infancy. There was no native Java support. Domino could render to a
web browser, but AJAX was totally unheard of. You didn't have
embedded views, or editors, or layers, or alternating rows colors
in views. I'm not sure there were Image Resources that far back.
Certainly there was no XML processing.
I really want to know who this guy worked for. I need to go make a
sales call.
3 Karl-Henry Martinsson Permalink @Nathan: The posting IP was 194.51.44.2 and a IP lookup gives gw.tokheim.com. Going to www.tokheim.com shows it is a manufacturing company, so it seems like he posted from work.
Intersting fact: "tok" in Swedish means "crazy (person)" or "fool".
:-)
And "heim" in German (and "hem" in Swedish) of course means "home".
Would you like to work for a company called "home of the crazies"?
;-)
4 Karl-Henry Martinsson Permalink @Nathan: By the way, I wrote a editorial system in the end of 1997 for Computer Sweden, the magazine I worked for before moving to the US. It was still in use last summer, surviving at least 2 (of not more) attempts to replace it with a "professional" (purchased) system. Sure, it still looks like it was built in 4.6, but it was (is?) fast and efficient. There is not much complicated code, the main code is in the export functions to generate Quark Express files, and I am sure that part has changed when they switched editing platforms. They been doing some small modifications/adjustments, but it was (is?) largely unchanged since
1997. Now they are running Notes 7, AFAIK.
But otherwise I agree, more complex applications need to be
refactored every few years.
5 Frank Paolino Permalink I think "Notes Floats My Boat" therefore I am (a Notes lover).
I have even made a screensaver of it that can be downloaded
at:
Notes
Floats My Boat
6 Yancy Lent Permalink This would be a fun Lotusphere mini event. The setup could be just like McCain's medical records review. For 4 hours interested developers can view this database(s) on lab machines. No phones (camera), no usb drives allowed in the room; nothing that could potentially copy the code. Only pen and paper to take notes. A birds of a feather would immediately follow to discuss.
@Nathan; I wonder sometimes how people come up with the "lines of
code" count of Notes databases. I hope they aren't basing it on a
count of the design synopsis. For grins I did one on the v7
catalog.nsf, pasted the Ctrl+a into notepad then word... 167 pages.
Thats a but load of meta.
7 Karl-Henry Martinsson Permalink @Yancy: That's a great idea. En extension of what Mary Beth did last Lotusphere, where we could bring in applications and the UI team looked at it and came with commenst on the design.
But to actually have developers look at an application and give
feedback on the design, that would be really neat. Of course, it
has to be retricted to a couple of applications. Why not have 2 or
3 problematic applications brought in, and they can be looked at by
different people?
A kind of "peer review". I think that would be very interesting, a
kind of "best practices/worst practices" on a real life
application.