Throwback Monday: Speech recognition and redaction

This week we’ll be having some fun and getting into the ‘wow it’s nearly a new decade’ vibe by looking at a few posts we wrote precisely this time 10 years ago, as well as some of the big stories of last Orange Rag newsletter of the past decade. How much will have changed and moved on?
First from the website, and on 9 December 2009 we wrote that Nuance’s new free cloud speech recognition app Dragon Dictation (based on Nuance’s Dragon NaturallySpeaking speech recognition technology) was now available on iPhone via iTunes Store.
Former editor and Orange Rag founder Charles Christian wrote: “The way it works is you dictate your words, which are then transmitted (via a WiFi or 3G link) to Nuance, where their servers perform the recognition processing and return the text to your iPhone. Text can then be sent as an SMS message, transferred to the iPhone’s Mail app or sent to the clipboard and pasted into an applicable iPhone app.
“After the text has been returned to your phone, you can correct transcription errors either by using a pop-up keyboard or by tapping an incorrect word and choosing the correct word from a list of suggested corrections that appears. Dragon Dictation requires iPhone OS 3.1 or later, and is currently free however according to Nuance this is a limited-time offer, suggesting they may introduce pricing if the demand takes off.”
Speech recognition was still new at the time – Google launched its voice search app in 2008 – but Dragon Dictation is going strong and Dragon NaturallySpeaking is what drives both BigHand and Philips’ speech recognition technology, which embeds it in a more user-friendly workflow.
Speech recognition obviously goes from strength to strength and at the start of this year, Gartner predicted that, by 2023, 25% of employee interactions with applications will be via voice. For a pretty cool graphic showing the timeline of speech recognition developments click here:
Also on this date a decade ago we flagged a new whitepaper from DocsCorp looking at how to properly make redactions in a PDF! Just as a word of warning, this looking back business could throw anything up.
Charles wrote: “The whitepaper makes the point that while redacting paper documents is a fairly simple process – get a marker and strike-through the text to redact it – redacting PDF documents is more complicated. PDF documents are constructed in layers. For example, text is on one layer and images on another. Thus, redacting text with opaque objects is not a foolproof method of redaction as you are simply adding another layer, which can be peeled back to reveal what is underneath. To add another layer of complexity to the redaction problem, you need to consider the underlying structure of the PDF.
“Depending on how the PDF was generated – ie a Word document converted to PDF, or a scanned document output as a PDF – there may be more than one layer of information that needs to be redacted.”
What’s amusing is how little has really changed. DocsCorp is still spreading the message that your redaction tool must burn out the text from the PDF to ensure it is completely removed from the document.
In a paper in April this year they said: “Your redaction software should be able to permanently remove text, graphics, and images from the PDF document. So, if someone copies the content into another application, the content will not display.”
Here is a link to that paper with handy top 10 tips: