I wrote the following for a Data Doctor's forum discussion on this topic. Sharing it here with small edits for a broader audience that might find it in search. TL;DR at the bottom.
The assertion that using present day country and country subdivision codes is a better approach than the more fuzzy historical names sacrifices accuracy and provenance for validity. Something can be valid, i.e., a meaning that is authoritatively consistent and uniformly understood, for accuracy, the property of being correct and exact. When representing a meaning, both are important. Using the current ISO codes and more locally accepted standards for subdivisions maximizes validity. I always want to use them. However, sacrificing accuracy for validity is similar to unjustly convicting a person in court for a just and honorable action in order to follow the letter of the law instead of the spirit of the law. Ideally, we achieve both accuracy and validity, but accuracy is clouded by the veil of decades or centuries of time in genealogy.
Unfortunately, ISO codes are time sensitive. They change on a regular basis. The ISO sends out updates when area boundaries change. We may see this soon for the Crimea, for example. That's why in sophisticated geocoding schemes, one must use the accepted standard code and a period of time (with a start and end timestamp, or a point within the period) in combination. If the ISO codes reached far enough back in time, they could be adequate to our purpose in genealogy using a code and a point in time. That's what businesses do in industries like property and casualty insurance. Unfortunately, we need to reach much further back into time than any business supported by ISO codes.
Using an example of Maienfeld bie Magdeburg from this thread, we have a choice between validity and accuracy. If we go validity, it seems we must choose to define the location as Magdeburg, Saxony-Anhault, Germany (assuming English place names for this discussion; natural language selection is a valid and interesting, but different, topic). In doing so, we have lost accuracy and to anyone without a direct knowledge of the provenance of the data, it's less accurate than Maienfeld bie Magdeburg, Duchy of Magdeburg, Saxony-Anhalt accompanied by the applicable point in time. If we have the latter information, we know it's maximally accurate. Since we have provenance due to using the original representation, anyone can do their own validation and primary research on the meaning. If we use the modern ISO standards, the accuracy and the provenance is lost. The possibility of coming to a better understand of the original meaning is lost because the original representation is lost.
In the example of a dated marriage certificate, the original name is quite important. A determination that Duchy of Magdeburg is the correct representation is only accurate if the ancestor was married before 1806. After Napoleon conquered Prussia in the War of the Fourth Coalition in 1806, the Duchy of Magdeburg was dissolved on 7 Jul 1807 by the Treaty of Tilsit. After that date, the accurate designation is probably Maienfeld bie Magdeburg, Kingdom of Westphalia, First French Empire. After 1814-1815, that was no longer accurate, as Napoleon and his empire fell in 1815. Prussia retook the city during the War of the Sixth Consolidation. Since Magdeburg is on the west side of the River Elbe, it didn't get absorbed into Prussia, which happened a long stone throw away over the river to the east. I didn't look at what it became on the west side of the river, but it wasn't part of Prussia.
Speaking of that long stone's throw, are we sure the location is called "Maienfeld bie Magdeburg" instead of "Maienfeld bei Magdeburg". "bei" in Deutsch can be translated "near", so it's the same as saying "Maien's Field near Magdeburg", or perhaps "May Field near Magdeburg". If we lose the original spelling, even that is a loss as it can't be reverse engineered from the modern geography standards.If Maienfeld was near but east of the river, it became part of Prussia in 1815. That distinction is important to our standardized names and codes today. Today, Magdeburg is one of three independent cities in the state of Saxony-Anhalt. But as far as we know because we have the original name, "May Field near Magdeburg" may be near the city, but east of the river. If so, the correct modern name is Some-Unknown-Location, Jerichower Land, Saxony-Anhalt, Germany. What would we call the location in this case? We don't have information to name it accurately and validly if we use modern names. If we chose to force a modern name, we lose accuracy and many clues to finding the actual location when we use the modern names.
Knowing the original designation can even help with the fuzzy dates we often have. It can also tell us something of the language, the religion, the politics, and the culture of the person who selected that name. All that is lost if you only use the representation Magdeburg, Saxony-Anhault, Germany.
Wrapping all that up, and changing the topic a bit, I suggest a place to look for what may be Maienfeld bei Magdeburg. If it transliterates to "May Field near Magdeburg", the place could be related to the Sack of Magdeburg on 20 May 1631. On this date, Magdeburg was destroyed by the Catholic League and the Imperialists at the direction of Ferdinand II. 80% of the population died and nearly 90% of the buildings burned. Looting continued for three days. The city was rubble until 1720 - 89 years! It was a big deal and it happened in MAY! It reminds me of the National September 11 Memorial and Museum in NYC, and makes me wonder if Maienfeld bei Magdeburg was originally a reference to a cemetery where the dead were buried, the field where the final artillery bombardment began, or the location of the final negotiation where the city refused to surrender after a two-month siege before Pappenheim sent in 40,000 troops to raze the city.
On a more peaceful note, it could be a reference to the tulip fields prominent at the 1600s and 1700s and today. The tulips start to show vibrant colors in May at this latitude. Perhaps the location is related to tulip fields.
In genealogy research, the last thing we want to do is obfuscate real meaning in the name of using a representation that's close to accurate, but assured to be inaccurate in many situation. Even when accurate, it reduces the trust the genealogist has that the location is accurate. As you can see by the attempt to research the location of Maienfeld bei Magdeburg, the actual original name carried a lot of embedded meaning to help us consider pieces of the puzzle. The modern name without the context might lead us to only look west of the Elbe when it may be to the east in the county of Jerichower Land.
TL;DR
The meaning is more important than the way we choose to represent it. If we haven't lost meaning, we can accurately represent it in numerous ways. If we lose meaning for sake of the way we represent the meaning, we've compromised the purpose of the representation. That's why I personally continue to assert that use of the name of a place as it was on the date of an event (birth, death, residency, etc.) is the better option when only one representation is available. The modern representation can often be derived with the original representation, but translation the back to the original is often not possible.
It would be really nice to have a second data point for each location - one for present day name (optional) and one for the name at the time of the relevant event, e.g., birth, date, residence.
I hope you'll consider this point of view.