What it should say is if you are running an event that is dual-rated or regular rated the recommended delay is 5 seconds, and if it is quick rated only then the recommended delay is 3 seconds.
The folks running G/25;d/5 do so to ensure their games affect regular ratings too.
Historically, 30 minute games were regular ratable before dual rating existed, and only games faster than 30 minutes were quick-rated. In 2012 we changed how time controls were parsed and that’s when the MM+SS rule was codified so MM+SS between 30 and 65 became the definition of dual-rated. There were probably G/25;d5 games rated prior to 2012, but the time control field was not being parsed and standardized so finding those events would be challenging.
In 2000, when dual-rating was introduced, it was to increase the number of quick-rated games in the ratings pool, so it did not impact the number of regular-rated games being rated.
Blitz games were split off from quick-only games in 2013.
I think that most only care that they are regular rated. 
I suspect if we got rid of dual ratings (and the Ratings Committee has a proposal to deal with staleness that may make dual ratings superfluous), nobody would notice.
If you haven’t watched the video of the Ratings Committee’s online workshop, I highly recommend it, a lot of time was spent explaining the staleness proposal.
I think with any variant of that staleness proposal we’d need some way to get that “will be rated as X pre-event” number from US Chess for pairing, rating estimation, performance calculations, etc.
I can understand and give the two sentence explanation but in terms of numbers I’d need a spreadsheet and time to work it out in practice
I don’t think we really know what the impact and usefulness of this method is yet.
What gets messy is that this will impact all 6 ratings systems, and adding 6 ‘experimental’ ratings systems so that we can study the impact on each of those ratings systems could be an implementation challenge.
I’m assuming a high percentage of the events played at G/25;d5 are scholastic events. With delay being mainly just a US thing, what kind of time control is used in scholastic tournaments in other countries?
It’s also interesting that the most common blitz time control includes an increment but the most common quick-only time control has no increment or delay.
I don’t know where to go for information on time controls common in other countries, but the US Open is good place to ask that question of many people.
There may be some regional influence on what time controls are used, one organizer running a lot of events could raise that time control in the relative rankings.
