Who's afraid of Text Validation?


Have you been sitting with a potential client demoing a form and the handwriting recognition fails? This moment is by far the easiest time to sell digital writing to someone and it might be your last chance!!  

There are a number of ways you can optimize your demo forms prior to getting in front of a prospect. As with all good things preparation is everything. Here are a few tips to use to get the Validation ball back in your court.

Demo Forms

Make sure to build your demo form with fields that is easy to recognize: First names, last names (make sure that test your and your clients names exist in the lexicon before the demo), country of residency (not nationality, the built in resource recognizes countries, not nationality, for example Sweden, not Swedish). It is also usually good to demonstrate something that relates to your client, like a product group, models or similar suitable.

Don’t use fields that are hard for the recognition engine like emails or difficult product numbers.

handwriting examples numbers and letters

Standard dictionary is another resource that never should be used when you are working with forms. It is many (many!!!) times to slower in terms of performance and the power of this resource comes into play when you recognize sentences, not single words.

It has logic for using other words in the sentence to pick the correct one, take a look at these two examples:

“The house is on fire” and “The house is on five”.


Fire and five generally looks very similar when you write it, but to us it is obvious which one is the correct one.

It is logic like this that makes the Standard Dictionary powerful because it knows this as well.

Resource types


It requires a bit of time to optimize your form, but it is simpler than you think. When you know the set of words (set of characters) that will be entered in a field you should use something XMS Penvision call a lexicon. This is just what it sounds like; you provide the MyScript recognition engine with a set of entries that it should pick from, nothing else. The recognition engine will do everything (and a bit more) to find a match within this set (It will return an empty value if it is totally out of hand).

Creating one of these lexicons is done in minutes, what takes a bit of time can be to actually collect the data. I am going to use car brands as an example here as an example. What you do is to create a text file with one entry on each line like this:


Example file

Log in to the XMS Penvision Partner Portal and find Custom Lexicons under the Order menu.

Portal login page

Select lexicon and upload your file:

custom format location in Anoto Live Forms portal

Custom formats

You will receive a file back that you upload to DesignTool:

location of add custom file in design tool

Example file (unzip before use)

Pendata has built a range of validation resource tools for various applications. If you need any assistance with yours get in touch and tell us what you need!

Gustaf Byström

Written by

I am Gustaf and I am co-owner and CTO. I am very passionate about running highly efficient and agile development teams.

Leave a Comment