If it were true that we could test features in isolation such that we could have a relatively high level of confidence in those rollouts with relatively little testing, then sure. But that isn’t the case. We have only vague ideas of what have changed because the release documents are frankly awful (and there have been many times they just don’t put stuff in the docs at all, so you just get to be surprised by the change). And then we have layers of customizations on top which of course are not tested by Epicor, and we have no way of knowing if they will be impacted or not.
So the only way to test is to test literally everything. I do not have time for that even once a year let alone twice. The tradeoff you are imagining where the releases are simply smaller and less painful doesn’t exist. In the reality in which we live today, which is that Epicor’s upgrades will always break a ton of stuff, and the only way to find it is by full end to end testing, twice a year is not realistic for them or us.
Alisa. I know you’ve already voted for the DevOps idea. This is what it is all about. We don’t live in that world, but I would love to.
So, what can we do to get there? Less upgrades won’t fix it. The only way to test everything is through automation. That was impossible with the smart client. With the browser and APIs, we may actually have a chance. Other companies are doing it, why not us?
… and you didn’t even use the term “tech debt” once! Impressive.
This process is the poster child for our own tech debt (not having a fully implementable testing script) AND Epicor’s tech debt (moving fast and breaking things). The software isn’t inherently buggier now, but there are more pieces to it and more of us using it.
interesting observing this dialog and others going around the EpiUniverse… some examples I have read/heard here and other places
one argument is fewer releases (once per year)
have more features (Epicor ideas are coming in fast (110 over the last month)
talk of CICD - continual releasing of small features
Feature flags… releasing items with feature flags that allow them to be implemented by choise
Kubernetes -
Delay releases
Allow patch back of features to older releases
Move faster
Move slower
Make it all browser compatible
Keep the smart client alive until we get everything moved
etc etc etc
Many of these dialogs disagree with each other and/or support each other. I would LOVE to answer every Epicor Idea with a new feature, but that would not be possible or desired. Choosing the important ones is a challenge I have accepted. This is fun!
yeah, I guess you got me there… but as annoying as the UX bugs are, it isn’t like the Smart Client didn’t break pretty often. I spent a LOT of time after every upgrade moving fields around and fixing screen issues.
True, I didn’t implement new customers into the Kinetic UI until 2022.2, and they’re still moving things around. And we can rest assured that when the NEXT big thing comes, we’ll be in this same boat all over again.
Consider it Tech Support Job Security… and WE’RE Tech Support. (I’ve been saying that ever since I WAS Tech Support (not at Epicor) lo those many years ago…)
I think if you need that level of control (like we do) that’s what On Prem provides. I feel for you SaaS folks and I know Epicor is doing their best to minimize disruptions hang in there.
I feel for us on-prem folks! We’re getting enough rope to hang ourselves! We make customizations that are difficult to upgrade, and boy we do! What percentage of us on-premers are constantly slipping into sustaining support? And I get it. Our world is changing. All. The. Time. We just want it to stop so we can catch our breath. We just want a little control.
With the move to Kinetic has been a journey for all of us.
What hasn’t been transparent from Epicor is the features that has made the Classic software great to use and how it has been implemented in Kinetic.
Export to Excel - Status working similar
Paste insert and Past update - Status - works on some grids, not on all
AP Invoice payment process - Status not sure, I did see in the past that paying multiple invoices was difficult in Kinetic
Publish and Subscribe with the Open with command - appears to be working in 2023.2 now, not sure if across the board.
etc…
To help adoption and for upgrades, if Epicor would share by module what’s is working (or not working) so users can make a educated choice.
New features are neat - but if basic functionality isn’t addressed, it’s step backwards.
If Epicor can provide this status, maybe this forum could provide the means to report that status.
I think we all want the same things - fewer bugs, more stability, less disruption, oh and yes new features too. We just disagree on the approach to get there. On the one hand you have the eternal optimists (Mark) who want Epicor to get on board with a true dev ops/cicd/feature flagging mindset where we can ABU (“always be upgrading”) with minimal disruption. On the other hand you have the eternal pessimists (me) who do not believe Epicor will ever be able achieve this level in the development organization, accept that the upgrades will always be painful, and prefer to minimize the disruption by just doing fewer of them.
Ha! I just popped in here on the weekend to see what to expect. I got busy with end of month stuff and missed this thread. I’m happy to see it.
One factor I didn’t see mentioned yet - Scheduling a major software upgrade on DST weekend was a, uh, bold choice. I’m glad they’ve pushed out the rollout so there’s a real world chance to see how date and time handling goes.
As a cloud customer I would not change the cadence. I would worry if we went to once a year we would feel like we would be waiting forever for fixes and new features.
As far as testing, every patch or version change I prioritize like this: I don’t wany any front line employee’s calling me when we go live. I have a saying for management I’m focusing on making your people happy first which should make you happy. I print every daily document such as packing slip, BOL, Invoice, Traveler, Purchase Order, Quote, Quote Worksheet, Check’s, etc… These will find any show stoppers, the next level I print a Time Phase, Shop Schedule, Inventory List, Order Backlog, Open PO Report. This list is both for testing and emergency documents I print weekly to use as a contingency plan if Epicor is unavailable for any reason.
Finally I go thru a couple dozen of my custom dashboards which gather data all over the place so if a field or something has changed this process will probable find it.
We have been trying to fix and communicate updates, but sometimes communications of this level are difficult. but in summary, here is the status of each of the items you pointed out:
Export to Excel - Fully featured just like before… and in some cases, it works better. You can EXPORT even from landing pages (see note below on PASTE).
Paste insert and Paste update - This is tricky… we have two types of grids… “Landing pages” and true grids. Paste does not work in landing pages. We are still working out challenges with landing pages. For example you used to be able to update ABC Codes on the grid view, but now you cannot do that. BUT for things like GL Journals, Order Details, Order releases, Job Details, Mtl listing in eng workbench/job/quote (and many many more), you CAN paste update/insert into the grid just like before (it is a true grid).
AP Invoice payment process - I believe that this was all fixed in 2023.1 If not, we need to know about it.
Publish and Subscribe with the Open with command - TOTALLY fixed!
I tend to be a optimist (like @Mark_Wonsil) believing and trusting in the process… I have seen vast improvements in our process and planning over the past year+. We are listening now more than ever. Our developers have been working around the clock to resolve the very complex issues revolving around migrating a complete software package from an on-prem/client based system to a cloud based/browser based system ALL while keeping both UI’s intact.
A Year ago, we pivoted our emphases towards making the UI/UX experience the best you can get. This does require some fixes where it might have been broken, but also required some updates to how our screens look and feel to make it “feel” much better (this goes beyond bugs). Some of these changes have made the browser based UI better than the old smart client. We are not done. I am optimistic about our direction and progress, as well as our future direction.
Thank you ALL for your comments on this and other topics on this forum. It does help us understand how to serve you better.
Still haven’t seen any formal announcement from Epicor about why the upgrade was postponed, but I do now see a date out there for .6. I am very curious to read those release notes and see what is getting fixed in that release since I presume whatever is there would have been the reason for the delay.
I start with first downloading the clients on certain servers/stations.
Making sure the EDI process still runs correctly since it is customized.
Verify all customizations and personalizations
Force Validation (no Check) all customizations and personalizations
Recompile BPM’s
Check Configurators
Check to be sure the MES windows pop up with our customized window.
Redeploy all Dashboards
Asking Customer Service to input some orders that will make testing everything (Quote to Cash) complete.
Run MRP
Ask Schedulers to do their thing for the orders that were entered for testing.
Run MRP
Go through our whole MES processes.
Test our Report Quantity process.
Test out our label printing process.
Test out our Packout procedure.
Test out our Master Pack procedure.
Test out our Invoicing procedure.
Make sure all forms print correctly.
Ask Engineering to enter some parts to make sure Engineering Workbench works properly.
On major upgrades I ask Accounting to make sure all AR and AP Groups are posted as well as GL Journal Entries and make sure all Review Journals are posted and run all their reports so they know the status of their Accounting system.
Alisa. I know you’ve already voted for the DevOps idea. This is what it is all about. We don’t live in that world, but I would love to.
We make customizations that are difficult to upgrade, and boy we do! What percentage of us on-premers are constantly slipping into sustaining support? 
And I get it. Our world is changing. All. The. Time. We just want it to stop so we can catch our breath. We just want a little control.
