Re: [EXTERNAL] Does every DB record need to produce a PV?

[Simon Rose via Tech-talk <tech-talk at aps.anl.gov>]

It should also be noted that what counts as an "internal PV" can highly depend. But what we usually think of at ESS of an "internal PV" is an intermediate PV that is used to calculate, to trigger, or some other processing that is only of interest to the IOC itself and is clearly not a part of it's "public API". I suppose that it is possible that you might want one of these to no longer be regarded as internal, but more likely a new record would be created that serves whatever purpose you are trying to solve.

Cheers,

Simon Rose

ESS

---

From: Tech-talk <tech-talk-bounces at aps.anl.gov> on behalf of Anders Lindh Olsson via Tech-talk <tech-talk at aps.anl.gov>
Sent: 18 February 2026 16:23
To: Pearson, Matthew <pearsonmr at ornl.gov>; EPICS Tech Talk <tech-talk at aps.anl.gov>; Ralph Lange <ralph.lange at gmx.de>
Subject: Re: [EXTERNAL] Does every DB record need to produce a PV?

 

> How do you handle database templates from upstream packages like areaDetector?

Depends on the case, but generally falls into one out of three solutions:

• Include the character in the macro for the "root" of the record name (e.g. R="foo:#")
• Patch upstream (we have a built-in patching mechanism in our EPICS setup)
• Maintain site specific db/template/substitutions files

> I can see where that might be useful, but how do you handle moving PV names from internal to public or visa versa? Do you add a public alias name, or change the record name?

This is a completely different problem than internal/public, but the way we handle renames for things which have been in production and/or (potentially) has consumers is carefully (if at all).

Normally goes like:

1. Rename* + set up alias for old name
2. Fix or append archival (depends)
3. Update everything else when there is time (OPIs, alarms, etc.)
4. Remove backwards compatible alias some time later (not sure we make it to this point).

* I always struggle with the idea to rename records as there is no persistence or unique identifier outside of the name.

Cheers

A

---

From: Pearson, Matthew <pearsonmr at ornl.gov>
Sent: 18 February 2026 15:47
To: Anders Lindh Olsson <anders.lindholsson at ess.eu>; EPICS Tech Talk <tech-talk at aps.anl.gov>; Ralph Lange <ralph.lange at gmx.de>
Subject: RE: [EXTERNAL] Does every DB record need to produce a PV?

 

Hi,

 

I’ve never found a need to mark some record names as internal and some not. I can see where that might be useful, but how do you handle moving PV names from internal to public or visa versa? Do you add a public alias name, or change the record name? How do you handle database templates from upstream packages like areaDetector?

 

Cheers,

Matt

 

From: Tech-talk <tech-talk-bounces at aps.anl.gov> On Behalf Of Anders Lindh Olsson via Tech-talk
Sent: Wednesday, February 18, 2026 5:43 AM
To: EPICS Tech Talk <tech-talk at aps.anl.gov>; Ralph Lange <ralph.lange at gmx.de>
Subject: Re: [EXTERNAL] Does every DB record need to produce a PV?

 

> Do other sites using the Channel Finder have similar issues with internal PVs being included in its PV name database and making it harder to identify the public ones? Do you filter out the internal PVs at all? We do not have any real issues

> Do other sites using the Channel Finder have similar issues with internal PVs being included in its PV name database and making it harder to identify the public ones? Do you filter out the internal PVs at all?

 

We do not have any real issues at ESS, with ca 14 million between production and lab networks, all in ChannelFinder. We have talked about filtering out internal PVs but do not and most likely will not. We instead just mark the internal ones (using a pound symbol in the record names).

 

 

Cheers

A

---

From: Tech-talk <tech-talk-bounces at aps.anl.gov> on behalf of Ralph Lange via Tech-talk <tech-talk at aps.anl.gov>
Sent: 18 February 2026 09:48
To: EPICS Tech Talk <tech-talk at aps.anl.gov>
Subject: Re: [EXTERNAL] Does every DB record need to produce a PV?

 

On Tue, 17 Feb 2026 at 20:19, Sukhanov, Andrei via Tech-talk <tech-talk at aps.anl.gov> wrote:

From my perspective, an EPICS IOC with ~2000 PVs lacking descriptions, for a device without a Programmer’s Manual, is unmanageable.

 

Certainly. I completely agree. I've come across such things, and they're a pain.

But that's bad application quality, which is a different thing.

 

Like reading a bad book, which can be a painful experience. Where I don't think Microsoft Word (the tool they used) nor Merriam Webster (where they got the words from) is at fault.

 

Of course, nowadays, AI can write wonderful books. Amazon's ebooks section is full of them.

 

Cheers,
~Ralph