I thought it would be interesting to have many people’s perspective on Sampler and Starter Solutions.
My experience with it:
About 15 months ago, I used a Starter solution - Estimates - to demo to a client how FileMaker could be a much better solution for his needs than his current combo use of Excel and Word. It had taken me only a few hours of my weekend to “personalise” (as opposed to customise) the solution. It would have been even faster if I had realised I did not need to translate the UI because there was already a French equivalent available in the extras, which I don’t normally see since I work in English. When he saw how the solution was dramatically reducing the time he would spend producing an estimate, he was “sold”.
Then came the time to actually “develop” the solution. I had heard many comments of people about these solutions. One particularly marking for me was a fellow French Canadian who had done something similar to demo a project to City of Montreal, using an assemblage of sampler solutions. The project was accepted and she spent countless hours altering and harmonizing the structure and the look and feel of the contraption adding the features requested by the client. When asked she would clearly state that she would never start from a starter solution again.
So I started from scratch. Quickly, however, being a beginner, I realized that it was taking me too much time, as I had to research every single thing I was trying to do and, try out various approaches to understand them and figure out which one to use. Since I had to deliver something, I decided to put the solution “shell” (mostly theme and structure) I had achieved until then and went back to using the sampler demo and started customising it.
I was able to deliver something within a month. (Thank you @Thierry Guemboura for countless hours of sustained assistance!)
My client was ecstatic but quickly realised the greater potential (I had seen it while developing but did not want to add more to my ordeal) of his new tool and asked for version 2.
Since the sampler was getting more and more difficult to customise, du to its original structure, I decided to rebuild from ground up. I was progressing in baby steps and really busy with other urgent matter, I had to halt working on it for various long streches. My client started to get impatient (won’t blame him there!) So I decided to do a “patch” to the production file with some features he was pressing for.
I downloaded the prod version and started developing on it. Quickly the limitations of the structure showed their face and I spent an outrageous amount of time just trying to adapt the structure to be able to add the new features. It was starting to take too long again. So instead of a “patch” I downloaded another version of the prod and did a few changes to fix certain bugs and change certain calculations to reflect new grouping requests. Reuploaded.
Continued to work on the “patch”. New series of “urgent” change requests. This time, I worked during off hours directly on the prod.
At this point I have 3 versions, the prod version, the dev version (the sampler modified one) and the rebuild that is far behind.
Each version has important elements the others do not have. The dev version is by far the closest to the end result desired but it does not have the proper calculations and field definitions I integrated to the prod version. It would have been wise for me to document the changes I made. In fact I started doing it. But because I often have to try various ways until I find the right one, my notes become unreadable at some point.
After trying to compare xml with Altova Diff Dog, trying to compare DDRs in 2 browsers windows (the elements are not sorted in the same order, making it very hard to follow), print screens of the fields list in the database manager to compare with a live one, I was pretty discouraged with how hard it was to get a list of the differences.
I finally printed the 2 ddr in .pdf and annotated dev ddr with the differences shown in prod ddr and then went to alter the dev one. This is where I am now.
Given my experience with Sampler and Starter solutions, I have really questioned why FileMaker put them out there in the first place, especially since most veteran will tell you to never start with a Starter solution.
I felt it was a disservice to the newbies. For example the navigation techniques in the Estimates is very hard to understand at first. If you try to alter the number of layouts included in the navigation system of the solution, you break everything.
There are ZERO comments on the comment lines. So you have no idea what the fields do and why they do it unless you spend countless hours trying to figure it out.
Some practices are… questionable to say the least. For example, still in the Estimates starter, the match field between Estimate Lines and Product is Article, (the item name basically) not an ID. That limitates a lot what you can do. Yes you can add an ID and fid and such… but you first need to know how to do this correctly.
I was pretty much frustrated and against them for a while but I have changed my mind. I think they do have their role to play, but such role should be clarified or they should be improved, if they are to be presented as valid starting points.
Pros (use for):
- A quick demo to entice a potential client
- A quick starting point to have “something” to teach specific techniques
- A learning tool to deconstruct, play with various features (e.g. layouts)
- A learning tool to learn Data Viewer and Debugger (provided the bugs have been added so the learner has something to work with)
- A quick base to do systematic tests to learn how the platform behaves
- A tool to create themes
Cons (To not use for):
- Base of a solution
- Extension to a solution
- Solution (except maybe if it is for personal use AS IS e.g. address book, cd collection)
That is my take on it… Have you used them? What for? How did it turn out? What would you recommand?