The PIC32 is a sad news.
Professionally I use PIC for 15 years. PIC16, PIC18 with external MAC/PHY, PIC18 with internal 10Meg MAC/PHY, PIC32MX and now PIC32MZ.
The last almost good combo was : PIC32MX / MPLAB IDE / MLA (Libraries, TCP/IP, USB, ..)
After this all launch from Microchip is "catastrophic" : deprecated the old one before the new are operational
* PIC32MZ "EC" : chip full of bug, end of life after 1 or 2 years ! now replace by PIC32MZ "EF"
* MPLAB X : new IDE based on Netbeans, multiple years before somethings usable (still not accept space in filename in 2017 !)
* Harmony : new "middleware"/"bloatware", mandatory for USB or Ethernet because the hardware is not well documente. You can read this topic "Harmony Hell in a Hand Basket Rant." on the Microchip forum : http://www.microchip.com/forums/FindPost/988424
Microchip is a very "closed" company, no discussion, hidden doc and source, compiler cost. The compiler is base on GCC but the "free" version is voluntary deoptimized, you have to pay 931€ (http://www.microchipdirect.com/ProductSearch.aspx?Keywords=SW006023) in order to have a decent compiler
Last but not least : Microchip bought Atmel early 2016, where is the coherence ? Microchip use Mips, Atmel use ARM. What future for MPLAB IDE vs ATMEL Studio ?
Microchip is incapable of making good supports and tools for ONE line of product, what happens with TWO different technologies ?
At work we are looking to escape from Microchip but for a small company and a lot product and history (15+ years) it's very difficult. My escape line was IoT and ARM based solutions with a lot of chips suppliers but this news close this one.
This day is definitively sad