During an interesting conversation with a senior buy-side participant a few interesting points were made which merit repeating here...
In many buy-sides the OMS was brought in for the equities teams to use and generally most buy-side OMS platforms work well in that space.  Over time the feature set has been rounded out to include some fixed income functionality but within the average OMS there is no fixed income trading workflow - so no native RFS/RFQ/countering workflow.  Instead the OMS stages out by FIX to a specialised fixed income system such as Tradeweb, Bloomberg, MarketAxess etc...
At the same time, since most buy-side OMS platforms had poor Fixed Income modelling capabilities many buy-sides built in house systems or bought another platform to offer Fixed Income portfolio managers something to use that made sense for their workflows.
Fast forward a few years and we now see many Fixed Income shops with a well rounded in-house built portfolio modelling environment which has, in some cases, been enhanced to add compliance functionality.  The Fixed Income modelling system is then integrated to a vendor OMS and the OMS integrates with the Fixed Income execution platforms.
Here's the question though - why integrate to a fully featured OMS when it provides very little real value?  And hence the title of "The NoOMS".  Perhaps in the case of Fixed Income buy-sides there is no real need for a vendor supplied OMS when a simple middleware based integration between the portfolio modelling system and the execution platforms could provide a low-cost, low-risk alternative?