Patent prosecution strategy has shifted in the last few years, but I think a prudent strategy may be to write a very good specification, claim broadly, and use divisional applications as an insurance policy.
In the Old Days, (prior to Festo), there was a general strategy of claiming very broadly in the first iteration of a patent application. The examiner would whittle away at the claims and the applicant would argue back and forth until a compromise was reached.
Then came Festo.
Festo stood for the general proposition that if the applicant gave away anything during the prosecution by amending the claims, the claims would be limited to only the embodiments described in the specification. The Doctrine of Equivalents for claim interpretation was (allegedly) dead. Immediately after Festo, there seemed to be a lot of hand wringing and doomsday scenarios floating around. Several different and rather outrageous strategies emerged, including filing zillions of claims at the start of prosecution with every conceivable combination of limitations so that claims could be cancelled or withdrawn, but never have to be amended.
Post-Festo, the overall strategy was to go to the Patent Office with a rather narrow set of claims and attempt to work your way broader. If you could get a narrow claim set allowed, you could file a continuation application and go with a much broader claim set.
The Patent Office has begun to put the brakes on this strategy. First off, they hiked the fees for extra dependent claims from $9 per claim in 2000 to $25 per claim for small entity, and from $18 to $50 for large entities. Extra independent claims used to be $38 and now are $100 for small entities and from $79 to $200 for large.
The second thing is that the Patent Office is proposing to crack down on the number of continuation applications. While these rule changes have not gone into effect, the change in prosecution strategy is quite drastic. No longer can an applicant go for narrow claims in hopes of avoiding Festo and then try to broaden the claims in continuations. From what I have seen of the proposed rule changes, a small number of RCE or continuations would be permitted, but any more would require a petition. I believe that the rule changes would permit unlimited number of divisional applications, however.
Given the current state of affairs, what is the preferred strategy? It seems prudent to me to go back to a broader claiming strategy, where you claim as much as you would like to get. Further, it makes a lot of sense to try to claim as many different inventions as possible in the original application, with hopes of getting restricted and filing a divisional application at a later date. The divisional applications may be very useful in the future if you run out of RCE or continuation applications.
To deal with the Festo issue, it also seems prudent to me to write as many embodiments and variations as possible in the specification. This is prudent regardless of Festo, but a well written specification with every conceivable embodiment should give the same coverage whether or not the Doctrine of Equivalents is applied.
One of the challenges with patent prosecution is that the application filed today may not be taken up for examination for another five years. The rules and court cases that guide our strategies today will have certainly changed, but it is critical to our clients that we use the best strategies given what we know at this time.