An exclusive gaming industry community targeted
to, and designed for Professionals, Businesses
and Students in the sectors and industries
of Gaming, New Media and the Web, all closely
related with it's Business and Industry.
A Rich content driven service including articles,
contributed discussion, news, reviews, networking, downloads,
and debate.
We strive to cater for cultural influencers,
technology decision makers, early adopters and business leaders in the gaming industry.
A medium to share your or contribute your ideas,
experiences, questions and point of view or network
with other colleagues here at iVirtua Community.
Four top tips
I was talking with a fellow IT professional recentlywho was saying that you should "program simply" but "design withcomplexity". I do understand the point that he was trying to make -that is, during the design stage take into account the problems in yourdomain so that these have been worked out as much as possible beforethe coding starts.
In many ways this is sound advice as otherwise a solution may beproposed that is inappropriate, inefficient or impossible to implement.However, if this is taken to the extreme, then the design may take onthe importance of the final deliverable in the minds of the designers.It may also become complex enough to be worthless!
document.write('\x3Cscript src="http://ad.uk.doubleclick.net/adj/reg.developer.4159/lifecycle;'+RegExCats+GetVCs()+'chl=plan;pid='+RegId+';'+RegKW+'maid='+maid+';test='+test+';pf='+RegPF+';dcove=d;sz=336x280;tile=3;ord=' + rand + '?" type="text/javascript">\x3C\/script>'); <ahref="http://ad.uk.doubleclick.net/jump/reg.developer.4159/lifecycle;chl=plan;dcove=d;sz=336x280;tile=3;ord=XPakY0gD9jwAAHz4aNgAAAJg?"target="_blank"><imgsrc="http://ad.uk.doubleclick.net/ad/reg.developer.4159/lifecycle;chl=plan;dcove=d;sz=336x280;tile=3;ord=XPakY0gD9jwAAHz4aNgAAAJg?"width="336" height="280" border="0" alt="" />
Last timetime I argued to "Keep it Simple Stupid", or KISS, with regard tosoftware implementation. However, it is not just the software thatneeds to maintain its simplicity - it is also the design. That is notto say the design will be simplistic but rather that it should be assimple as it can be in order to fulfill its requirements.
This in itself sounds straight forward, but consider what I havesaid here: that the design should be "as simple as it can be in orderto fulfill its requirements" - that is a slightly odd phrase. What I amtrying to get at is that the design should be no more complex than isneeded to help to understand how the software will need to beconstructed.
This differs from the requirements of the software itself. Remember,a design is an abstraction of the software that will be written - it isnot the software itself. At the end of the day, it should not be ascomplex as that software - it's a simplified representation of it.
Clear models in any Java-oriented design process, modeling is a keyelement. An overly complex model may actually be less useful than asimple model anyway. Why? Because a complex model, with a large numberof classes, objects, relationships and use cases may actually obscurethe meaning originally intended. Worse, the density of elements in themodel may actually mean that developers get "model overload" and switchoff.
If this happens, then the formal model may be forgotten anddevelopers will work from their own interpretation of the requirements.This may result in the production of an implicit design, which is nomore than a set of mental models.
So, how can simplicity in design be achieved? Here are some practices that I believe can help:
Don't try to test designs
As I have said above, a design is an abstraction of the software tobe implemented. As such it is not possible to test a design to checkits completeness, its suitability or to validate its functionality. Sowe should not attempt to do so. If we need to validate part of a designthen prototypes can be built. If we need to explore some complexconcept then a proof-of-concept implementation can be created. Weshould not pretend that a design is something that can ever be as allencompassing or complete as the resulting software. Instead the testshould be: "Is it fit for purpose?" If it passes this test, then it hasmet its goals.
Use the simplest tools
Tools are another area in modeling that people get hung up on. Somemodeling tools tend to promote the creation of large complex models -this may be to do with the modelers creating models for their own sake.However, the use of a dedicated, and isolated modeling tool tends tohelp promote this. Personally I am a big fan of tools such as Omondo- the Eclipse plug-in that can create models from Java source code andsource code from Java models. I like such tools because they arelightweight, do not get in the way of the modeling process and tend torestrict the creation of very large, and potentially meaningless,models.
Use Patterns gently
Design patterns are good things, yes? Not always. Due to their verynature, if a design includes a number of interacting design patternsthe end result may become, at best, complex and, at worse, opaque andpotentially useless. However, I sometimes wonder whether some designerstry and show how clever they are by producing complex designs thatencompass numerous design patterns. As Scott Ambler has said: "Use design patterns gently!"
Design with others
Of course most designers do not start out with the aim of producingoverly complex designs and models. Actually achieving simplicity indesign is a much harder thing to succeed at, than to strive for. Oneway to keep things simple is by designing with others. In my experiencethis tends to help to things stay focussed, to ensure that the designsremain meaningful and that they are as clear as possible given theapplication. It is essentially the design equivalent of XP's LINK pair programming.
SummaryIn conclusion, it is important to note that I am not advocatingsimple design, but rather the goal should be to aim for the simplestdesign that fulfils its purpose. Thus we should aim for simplicity indesign just as we should in actual implementation.