The riots and some stuff about the Arlathan society is talked about in the novel "The Masked Empire" and you learn a lot about the Eluvians as well, i would suggest picking it up for a read,
And in the novel Asunder High Seeker Lambart was at one point a Tevinter Templar so you can learn some stuff about tevinter there.
Thanks, I'll look those novels up. The truth is, I tried reading the first novel, The Stolen Throne, and I just didn't find it good writing. Its style was boring most of the time, and I usually had to force myself to read it, because of the lore info. A matter of taste I guess. I hope I find the other novels better.
That's a typical case of "reality is unrealistic". The Siege of Candia, for example, lasted 21 years. There are other examples in history too. 6 years of siege doesn't look too much compared to them. If you are looking for a modern case of a very big city under siege, we have the famous Siege of Leningrad. Hell, the attackers, the Tevinter Imperium, are known to have a capital prepared to face any kind of siege, so why not the elves?
Re Candia, the siege was never completely closed for more than 28 months, which lead to the fortress' surrender, and even when it surrendered it was not starved out. The siege began in May 1648, then was lifted in the first months of 1649 and was not resumed until June 1649 and even then only lasted for another two months. From then on, the situation went from a siege to a loose blockade. Ottoman control of the sea was also not permanent throughout the period 1648-1669. As late as June 1667 it was possible for a French contingent of reinforcements to arrive at Candia, which tells you that the siege was never completely closed from the sea.
In my post I have said "A siege of a large city would have hardly lasted more than a few months, especially if the city was completely cut off from food and water supplies.". I emphasize the word "completely". This requirement makes your example of the siege of Candia invalid.
Re your "other examples", they are all examples of sieges which were not completed, or are too poorly documented for us to say anything definitive:
- Skipping Candia, Philadelphia (1378-1390) is too much lacking historical sources to make any comments on it anything more than an educated guess. We don't know the strength of the sieging/besieged forces, we don't know how long did the city spend in conditions of a closed siege, all we know actually is two dates.
- Ishiyama Honganji -- I am sure you have read the article before posting the link to it, so you have also read the sentences "Attempts to starve the fortress failed, as it was supplied from the sea by enemies of Nobunaga." and "In 1578, a fleet was finally able to cut supply lines to the castle. By 1580, the defenders had run out of food and ammunition and were forced to surrender." So again, the siege wasn't closed, and once it was, the ~30 000 strong garrison still held out for a remarkable two years.
- Thessaloniki was also not closed off from the sea, and the Venetians had no problems supplying it. What they did ultimately have problems with was manning it

- Drepana's siege was broken from the sea by Carthage relatively quickly and was only renewed by the Romans around the end of the war - 241 BC. It takes the Romans 5 years to build another fleet of 200 quinquiremes which was dispatched towards Drepana no sooner than "the beginning of the summer" (Polybius) of 242 BC.
- The Solovetsky Monastery siege was also not complete and the siege only ended as a result of threachery from whithin (it was not starved out). I checked the Russian Wikipedia article on the siege. It goes into much more detail than the English article, and it seems that as a rule the siege was broken during winter, because of the cold and because of insufficient forces, and then renewed in the warmer parts of the year.

- The Siege of Tripoli - I didn't research this one much, but this sentence from Wikipedia was enough to show me the siege was again more of a loose blockade than a real siege: "The nobles of the city, who had betrayed the city to the Franks by showing them how it was being resupplied with food, were executed in the crusader camp.". Apparently it was a practice to resupply the besieged city

- Harlech Castle was readily supplied by sea. Wikipedia - "After a month's siege, the small garrison surrendered on 14 August.". The reference given is "Taylor, Arnold (2007). Harlech Castle. Cardiff, UK: Cadw. ISBN 978-1-85760-257-9." So much for this "7 years" siege.

- The 6 years long siege of Xiangyang is actually legitimate, altough it was made possible by the accumulation of a great amount of supplies, the availability of drinking water, and the 200 000 civillian population must have had means of producing food.
- The last example remaining is the siege of Gibraltar 1799 - February 1783. This one is also the best documented. It was broken three times, with the garrison being reinforced and resupplied, before it was finally lifted.
Read all about it in Wikipedia.
Lastly, Leningrad, which you propose as an example is famous, among other things, with the fact that the siege was never closed. The city was resupplied over the Ladoga lake:
http://en.wikipedia...._the_defenders.
In conclusion, I guess I can agree to the version of a few-years-long siege of Arlathan on the condition that it was incomplete and/or broken from time to time. A well conducted and complete siege that would last more than a year though, I don't believe to be possible.
Please note I haven't intended this thread to do nitpicking on the DA setting. It's just a collection of questions which sparked my curiosity and wanted to see how other people imagine them, or things that seemed inconsistent the way I understood them.
As for the walls, if you have WoT, you can see in the timeline that 1819 years passed since the elves started retiring because of the Quickening (or whatever the real reason was) until the siege. In those long centuries there were wars among humans, and probably among elves too (as suggested by the tales about the Brecilian forest in DA:O). Even if Arflathan didn't have walls at first (also an assumption; who says elves didn't make war against one another too in the old days?), it's more than enough time to build new walls for the city.
Ok then, I guess they must have had fortifications.
Re the Tevinter discussion, I wouldn't go so far as to make parallels between the Roman society and Tevinter. All of my questions about it remain, and I'd love to get to chat about it with someone from Bioware who's relevant to the topic. What did their body of laws look like, did they have anything resembling checks and balances, who held the legislative authority, etc.
I don't agree with your remarks on Divine Right, but that's a different subject altogether, so I won't go into specifics.
Regarding the Jews, I said "a position similiar to that of Jews in medieval and early modern Europe". Similiar in the restrictions on holding public office, and owning land. I agree there are similarities between the freedmen and the liberati, and I think they (the liberati) look realistic in the context in which they are described.