(quotes taken from this thread but moved here, because discussion probably belongs here)
I think it's a matter of picking a line to draw, and there are a lot of options within a range for where a person chooses to draw the line.
Personally, I tend to draw the line at 15:00, so ANTR is, and TMOLS isn't. But I could see drawing the line at 14 or at 17 or whatever.
The line here is usually 20-ish (or almost), that's why TCOT is being considered. IT is also longer because of the hidden part at the end, it would be around the same length as Count otherwise.
But why make the completely arbitrary cutoff point at exactly 15:00? TMOLS is literally only 3 seconds shorter. Does that seriously make the difference between it being an epic or not? 15 (or 20) minutes is a nice round figure, but music doesn't work in absolute numbers. Saying 'everything over 15:00 is an epic' is a very systematic, and dare I say narrow, way of looking at music, because '15 minutes' (or any other fixed time limit) is meaningless in music in general. Furthermore, music can be played faster or slower than on studio recordings: the TMOLS we have on Systematic Chaos is just a 'version' of that song, the one released on Chaos in Motion is actually 15:22. Is it an epic now? See how choosing a time-limit doesn't really make sense?
Taking another Dream Theater example. The version of Beyond this Life is stretched to almost 20 minutes, about 80% longer than it's studio counterpart. It's extended by making the instrumental part longer, again creating a different version of the same song. Is it an epic now that it's suddenly 4 minutes over the 15:00 limit, instead of 3,5 minutes short?
Personally, I think that looking at song structure and how different parts contribute to a whole makes more sense in trying to define whether a song is an epic or not. There is no set definition to what an 'epic' is, but if we take some of the great prog epics of the past as examples (Close to the Edge, Hemispheres, Supper's Ready etc.), then to me Dream Theater has 4 epics (A Change of Seasons, Octavarium, Illumination Theory, A View From the Top of the World)* and a bunch of really long songs. The reason I list these four and no others, is that they go through different movements and completely change the atmosphere accordingly. They don't really have many repeating parts, they list a continuing story, although fragmented, or parts of a larger whole, rather than something literal. And if it is something literal, it needs to involve a struggle of some sorts; a narrative arc consisting of a beginning, a journey and a resolution. That said, I realise this conflicts with what I said before and the fact that there's no set definition of what an epic is, makes it really difficult to discuss this properly. Thinking about what I just wrote and why I wrote that, I realise that drawing the line anywhere is going to be arbitrary and a case could be made for other songs as well I suppose (for example Trial of Tears, TMOLS and A Nightmare to Remember), but drawing the line based on a completely unmusical feature like mere 'song length' is definitely not something I can get on board with, because ultimately that's about as meaningless as for example the song's placement on the album.
I should clarify: I think there are multiple ways of thinking about this, multiple ways of categorizing songs, partly because there are not bright, clear lines laid down by reality itself or anything like that. I'll quote another post of mine from the same thread that goes into another way I break down Dream Theater's long songs:
Another way I categorize them when I'm thinking about Dream Theater specifically (the 15:00 mark is one I use across all artists) is to think about "classes" of songs. I think DT's longest songs break down into a few categories:
Six Degrees of Inner Turbulence is in a category of its own, being basically an album-length song.
A Change of Seasons, Octavarium and In the Presence of Enemies are a category of multipart songs with a lot of variety that greatly exceed the 20-minute mark.
The Count of Tuscany, Illumination Theory and A View from the Top of the World are long songs with a little less variety and less clear division into sections than the four longer ones, which fall close to the 20-minute mark.
The Ministry of Lost Souls and A Nightmare to Remember are more like long versions of typical DT songs, both falling close to the 15-minute mark.
And then the next category, which has a lot of songs, would be the 12-14 minute class.
What I'm doing here is pointing both to length and to structural elements of the songs. There are multiple axes of measurement here, including length and a number of structural and musical factors. And there are gradations along those axes, not sharp divisions.
For example, let's take the new one, which I've put in the middle category. It's closer in length to A Change of Seasons than it is to A Nightmare to Remember, and it has a more clearly defined set of distinct sections, which make it seem like it belongs in a category with ACoS while ANTR belongs in another category. On the other hand, it has a lot of structural similarities to ANTR, including being divided into three sections with a lot of shared themes between them. This is unlike ACoS, where the sections are really distinct and there's not a lot of shared melodic content from one section to another. So this suggests that maybe all three belong in the same category, with a range within them along a number of axes but basically placing ANTR at one end and ACoS at the other.
But then, as you rightly point out, what's really the difference between ANTR and TMOLS? Both of them have the three major sections, and ANTR really isn't that much longer—it's closer to TMOLS in length than it is to AVFTTOW. And then from there, why can't we bring in Trial of Tears, which you can compare to TMOLS in much the same way as you can compare TMOLS to ANTR? The point is that there are no bright, clear lines, only gradations. And when you want to take a subset of songs to talk about them together, you have to draw a line in a place where it might be difficult to see a huge difference on either side of the line. It's sort of like how we have to draw a line to determine when teenagers are legally able to drive, even though there is not a lot of difference between 15 years, 10 months and 16.
I don't think that this works to establish a bright, clear line:
Personally, I think that looking at song structure and how different parts contribute to a whole makes more sense in trying to define whether a song is an epic or not. There is no set definition to what an 'epic' is, but if we take some of the great prog epics of the past as examples (Close to the Edge, Hemispheres, Supper's Ready etc.), then to me Dream Theater has 4 epics (A Change of Seasons, Octavarium, Illumination Theory, A View From the Top of the World)* and a bunch of really long songs. The reason I list these four and no others, is that they go through different movements and completely change the atmosphere accordingly. They don't really have many repeating parts, they list a continuing story, although fragmented, or parts of a larger whole, rather than something literal. And if it is something literal, it needs to involve a struggle of some sorts; a narrative arc consisting of a beginning, a journey and a resolution.
(Leaving aside the SDOIT debate) For one thing, why is A View from the Top of the World in and The Count of Tuscany out? The Count of Tuscany goes through different movements and completely changes the atmosphere accordingly just as much as A View does. It doesn't have many repeating parts. It lists a continuous story that involves a struggle with beginning, journey and resolution. Is it because there are no formally numbered movements? That seems like it would be even more problematic of a division than length—you'd have to exclude Alive Again by The Neal Morse Band from being an epic if that was the criterion, even though that's a 26-minute song with a really clear division into sections.
I'm not saying this is an invalid way of classifying songs, but I just don't think it's an inherently better one than the one I proposed earlier or even length.
Now, to address the cases in which I do use length, since you criticized that and I don't want to seem like I'm backing off of that entirely in favor of the division I quoted from my other post:
I think it makes sense to have a classification of a particular artist's songs and a different, more general system of classifying songs across artists. With a particular artist like Dream Theater, we can look at their style of songwriting and find that generally if they write a song longer than 20 minutes, it will have these distinct sections, and the sections will be less distinct as songs get shorter, and make a division based on that like the one I proposed earlier (or like the more limited one you proposed, though I would not use it myself due to the exclusion of SDOIT and TCOT).
For example, I have a different classification of Iron Maiden songs. They're much less proggy than DT, and they only have one song over 15 minutes. I could go into a similar sequence of multiple categories for them like I did above for DT, but if we just asked what I consider an "Iron Maiden epic," I would include a lot of songs that are like 9-12 minutes in length and no more divided into clear sections than a typical 9-12 minute DT song. But I would never call a 9-12 minute DT song a DT epic. Different artists call for different divisions because the set of songs is different and the style is different.
This is where my 15-minute rule comes in. There are times when I want to think about all the very long songs or epics that I'm familiar with, across a bunch of artists. In this case, there are so many differences in style and context that it would be difficult to get in and start sorting them out according to a very specific rule. Instead, I need to draw a numerical line. And 15 minutes makes sense to me, looking at the songs I'm familiar with. I want Close to the Edge on one side of the line and Trial of Tears on the other, because I'm trying to think about "songs like CttE" rather than "songs like ToT." At the boundary, there are questions you can ask, like "why should Graves by Caligula's Horse be included and In the Name of God be excluded?" and I wouldn't fight you if you said it should be 16 or 14, but 15 just captures all of the songs I want to talk about under that heading and not many of the ones I think of as maybe not belonging in that category.