I've reverted this in changeset/90029800.
I would still like to know why you originally made this change. Responding to comments from fellow mappers is an important part of contributing to OpenStreetMap, and is essential to attaining useful high-quality mapping.

I've returned this to highway=secondary in changeset/89919652, because I think that is much more appropriate. I wouldn't object to this being changed to secondary_link, and/or the reverse direction being changed to a matching classification.

Reverted in 90028115

I think "normal" in that sentence means "not a _link". The example given in the footnote showed a highway link using a pre-existing unclassified road to join onto a secondary road. The portion of the unclassified road involved was then tagged as a secondary road. (Both roads were changed from secondary to tertiary a couple of months ago, but the same principle still holds).

I think that statement is relating to the case of a simple link with no intermediate junctions, with the instruction being to not choose an arbitrary division point.
I think this point is relevant (from the highway_link page):
"A preexisting street used to connect two major highways is not a link. Instead it should be tagged as the normal highway it previously was, probably equivalent to the lower classification of the ones it connects. This also applies where otherwise minor streets connect highways at an interchange."
And an analogy from the motorway_link page:
"Any other roads used to connect the motorway to a non-motorway, which can also be used by non-motorway traffic, should be tagged as something less than motorway_link, usually the *_link of the non-motorway."

If this is considered to be a pre-existing street, then the wiki would say that 'highway=secondary' is correct. On the other hand, if this isn't a pre-existing street, but the A1307 were a motorway, then the wiki would say 'highway=secondary_link' is correct. I think the reality falls into neither case, so there isn't explicit guidance for this situation, but I don't think the classification should decrease if the classification of the A1307 were raised to a motorway.

So, to conclude, I'd say that either 'highway=secondary' or 'highway=secondary_link' is correct; I think I now favour the latter slightly more.

I think 'secondary' is probably best (though without the ref). (I'm basing this partly upon the various wiki pages for highway links).

I still don't think that works. Using "highway=construction" and "access" tags is ambiguous, and I would normally interpret this to be indicating the access arrangements once the road is open. I had a look at instances of this tagging elsewhere, and it looks like my interpretation is the most common one in the areas I found. If you want to indicate that the access is temporarily restricted, then I would just add appropriate access tags. Your changes also broke Graphopper's cycle routing.

I've mostly reverted your changes (in changeset/89936254) because I still think the tagging is wrong (and routing was visibly broken). It might be possible to improve the accuracy of the tags at each end of Histon Road, but the way you tried doing it doesn't quite work. Personally I'd favour just leaving it tagged as open because the short stretch of inaccuracy doesn't affect anywhere that someone would actually want to drive to or from.

Oops, I failed to add a comment.
Revert changesets 89635942, 89634946, 89579256, and add a slightly different tagging to Histon Road to indicate that the closure is temporary.

I disagree. And even if they should be the same, I think 'unclassified' is the wrong tag for the northbound sign, since it's main role is as a slip road from a motorway onto a secondary road.

I think there is a bus stop here. I don't know if there is a physical sign here, but route 606 (previously 206) to Impington Village College runs along this road northbound in the morning (on schooldays).

I looked at that page but failed to see the schedule completion date above the table. The table itself gives estimated end dates of Spring 2021 for Phase B+C and Summer 2021 for Phase D.

I suspected so. Thanks!

I think there are issues with most of the changes in this changeset.
1. The southbound carriageway through the junciton with Kings Hedges Road is not under construction, and is not being changed as part of these works (though there is a slight change to the kerb line further north). This stretch of road is also still open to cyclists, and to vehicles accessing the Darwin Green site.
2. The section immediately south of Kings Hedges Road is not a cycleway. It is secondary road that is temporarily closed to most traffic (but still open to cyclists and vehicles accessing Darwin Green - the access to Darwin Green is why I didn't add access=no tags here).
3. I think the direction=both tag you added to the traffic calming on Windsor Road is unecessary, and I can't find anything on the wiki to suggest that this tag needed to be added.
4. I don't think changes you made at the junction with Victoria Road are particularly accurate or helpful - they don't really match either the current or future routes for cyclists.
5. I have no idea where you got "approx July 2021" from - it isn't in the latest construction update I've seen. The scheduled completion dates are different for different phases of the work.

I think this best thing to do with this changeset is to revert it entirely, and then perhaps add more accurate tags to the bits of Histon Road that are temporarily oneway.

Why did you edit relation/1016554 in this changeset?

Can you explain why you have made this change? As far as I can tell, the existing tags were more accurate.
Also, your recent pattern of edits (involving bus lanes in seemingly random locations across the world) suggests that you may be making edits based on a systematic rule without sufficient understanding of the area being mapped. If so, then you should read the policy on mechanical edits on the wiki (osm.wiki/Automated_Edits_code_of_conduct).

I believe so - I remember seeing the 'end of shared pedestrian/cycle area' sign just round the corner on Fendon Road (south) a few days ago. I don't know how far they extend.

I'm not entirely sure what you mean. It isn't usually possible to delete or amend existing changesets (once they are closed), so the way to fix mistakes is to create a new changeset with the corrections. I think I've corrected Tom Scully Motors; if there are any other mistakes that I've missed, than I suggest you fix them yourself (in a new changeset).

I think you've added an excessive number of nodes in this changeset (and the other one in the same location). Most of the nodes you've added change the position of the way by less than 1m (and almost all by less than 2m), and some affect the alignment by less than 10cm (compared to straight lines ignoring those nodes). GPS traces won't be accurate to less than 1m, and imagery often isn't that accurate either - it can be distorted in various ways. It is also rarely possible (and never sensible) to try to identify the centreline of a road accurately to a precision of less than 10cm.

You should also add addr:street tags when tagging address (assuming the street name is part of the address - see addr=* for more details).
Also, if these are semi-detached or terraced houses, then I think a nicer way to map them is to put a node roughly in the middle of each home, and add the address tags to that node. (For a terraced house you can either do each one individually, or do the two ends and use interpolation). There are some good examples of address tagging for a variety of situations if you look at the area around Norfolk Street in Cambridge.

These are not housenumbers, so should not be tagged with addr:housenumber. I think this needs to reverted. (The added data seems to just be duplicating existing data, sometimes even duplicating the name of a different building.)
Have you made any other changes like this?