MarcEdit and the Windows XP Sunsetting conversation

I’m again thinking about the need to seriously think about Windows XP and MarcEdit’s continued support for the nearly 20 year old OS.  What is pushing this again is the integration work I’ve been doing with Alma.   None of this work will be available to XP users, in part, because ExLibris (and rightly) uses TLS 1.2 for their website security certificate.  This isn’t supported on XP (and never will be), so the connection cannot be made to the Alma service.  I have a feeling more and more of these kinds of problems are going to come up, and the way to “fix” this on my end is to migrate MarcEdit’s windows version (the Mac version already is) to the .NET 4.6 framework.  Right now, I’m using 4.0 as a base because it is supported on Windows XP SP 3, but I have to do a number of hacks to keep things working as new technology keeps coming available.  So, I think I’m ready to let Windows XP go, and say that it is time for the MarcEdit community to put this system behind us.  I’ll likely leave the last version of MarcEdit with XP support available for download, and maybe this change will be marked by a version change number (MarcEdit 7), but it really needs to happen.

At the same time, I’m sensitive to the fact that XP has lived so long because it has been the primary system used in a number of developing countries for a very long time.  This is part of the reason I wanted to let folks know that I’m planning to make this change, and give folks an opportunity to provide some feedback.  I’ll also be doing a few things between now and June.  I have a lot of log data, and I’ll be looking at these logs to identify the actual Windows XP use.  A quick glance at my website stats tell me that Windows XP is a minority operating system — but I want to dig deeper.  MarcEdit is run hundreds of thousands of times over the month (per the update logs), and I’ve tweaked the logging to preserve the users operating system (currently, I keep nothing but a general geographic location).  My hope is that libraries will have long ago let go of Windows XP, but I want to do some diligence to make sure this is the case.  If I find that there is still significant XP usage, I’d likely extend my date before dropping XP support from my development code branch, but usage will need to be pretty significant.  But this is also why I’m welcoming feedback.  I do want to be sensitive to the user community and try to make sure that no one gets left behind.

As of right now, my plan is to leave the last XP compatible version of 6.x available for download, but this would be the end of life version which would receive no further updates.  The benefits for the MarcEdit community is this will allow me to remove large sections of code that exist only because I’m dragging XP along, as well as an improved and leaned down installer.  Likewise, I’ll be able to take advantage of some new coding structures that will allow me to more easily do some semantic web integrations, as well as make it easier to support some of the 3rd party integrations that we see.  Additionally, it will bring the Windows codebase into sync with the Linux and MacOS codebases.  By default, both of those systems make use of the 4.6 framework, which would make testing much easier on my end as now I’m testing code across 10 operating systems and 4 different .NET framework versions.  Finally, since this will be a breaking change, i.e. MarcEdit 7 would be developed to work with Windows 7+ (though, I believe it would also work on Vista, but I would no-longer be formally testing for that OS given it’s historically low uptake) on the Windows-side, I would ensure that users could install the MarcEdit 6.x and MarcEdit 7.x programs-side by side.  This is the same approach that I took when moving from MarcEdit 5 to 6, and when I dropped support for Pre-XP Service Pack 3 machines (of which, there were still a handful).

In a couple of months, I’ll start detailing formal plans.  I should also have enough data (from log analysis and active logging of OS for the next 2 months) to have a good idea of what the impact might be and if the time table needs to be shifted.

I’m tentatively planning to make this change in Fall 2017, though I’ll very likely have a working build of the MarcEdit 7.x branch in early June.

–tr