I have had this question in my mind for quite some time.
What is the best way to go throughout history and read? Would it be the primary texts, written histories throughout the ages that are still published? Or read contemporary history books?
My trouble is that I am used to the idea of a canon, or the greatest books written within a discipline, and with history I can’t seem to find that, other than books like “ A history of historical writing” but even those authors listed do not cover every time in history.
So in that case I would go to contemporary texts? Which ones would I pick? How do I know one is better and holds more academic value than the other?
My trouble is that I am used to the idea of a canon, or the greatest books written within a discipline, and with history I can’t seem to find that
If there is one thing that the study of history tends to do, it is to explode the idea of "canonicity." Primary sources may be fixed works, but they are rarely authoritative. The historian always has to ask who is writing, why they are writing, whether the work has been translated or copied (and if so, why and by whom), how they know what they do and - increasingly - how what is written there squares with all the other sources of data, both archaeological and documentation-wise, which involve the period and events.
To an extent, the same thing is true for secondary and tertiary sources, especially textbooks. The folks that write those don't exist in isolation, and their work is usually informed by the state of scholarship at the time, their personal philosophy and experiences, and the sources they have available to them, many of which are probably lost (especially in the case of interviews with living people, but also sometimes documents and artifacts). A lot of books which were highly-regarded at the time and even well-written have simply aged out - the information in there (or at least the interpretation of the data) cannot be taken at face value any more, because the state of scholarship has moved on.
In pulp studies, one of my go-to examples for that kind of thing is Margaret Alice Murray. In the early 20th century, Murray was a respective Egyptologist with sidelines in anthropology. She received widespread popular acclaim for her book The Witch-Cult in Western Europe (1921), wrote the article on witches for the Encyclopedia Britannica expounding her theory, and a less-well-regarded follow-up The God of the Witches (1931). While criticized by many academics, Murray's thesis enjoyed great popular success, and had substantial influence on mid-century occult movements and the development of Wicca.
Murray's works there were important and influential, but also severely flawed (and, I'll say it, just flat-out wrong). I would not recommend reading Witch-Cult to anyone looking for actual insight into with European witch trials - but at the time that it was published and for decades thereafter, Murray was an important influence in pulp fiction and occultism. So I would recommend people read her if they want to understand where Lovecraft was coming from in "The Dreams in the Witch House" (1933) by H. P. Lovecraft, "The Salem Horror" (1937) by Henry Kuttner, or even Conjure Wife (1943) by Fritz Leiber - but with the caveat about what they're reading and why.
So getting back to your question...
So in that case I would go to contemporary texts? Which ones would I pick? How do I know one is better and holds more academic value than the other?
The important thing is usually to get an idea of the state of the scholarship - contemporary texts are generally more helpful in that regard than earlier ones - and if you're really interested in the subject, you start tracking down the sources they cite and evaluating the information from there. It can take quite a while to get a firm grip on any particular corner of a historical subject, especially if most of the papers and journals are old or obscure - which is why academic institutions are such a benefit to scholars.
And, of course, the state of scholarship moves - keeping "up to date" on a subject usually requires a bit more investment than a subscription to a particular journal - as new information becomes available, or new techniques and perspectives cause old information to be re-interpreted, a lot of the stuff you "learned" from those first history books might have to be unlearned, or at least caveated heavily. Being able to pick out those distinctions is a good hallmark for when your understanding of a subject has increased.
I like to use the reading lists from PhD comprehensive exams. There are a few programs who put sample ones online.
Wisconsin-Madison has one of the most comprehensive (and is also a great program).
Apparently Harvard recently put some of theirs online, and that's a pretty solid looking set of topics as well.
If you're interested in medieval, shoot me a PM (OP or anyone)--I'm happy to share mine, but it would be useful to talk about tailoring my 300 books to your interests. ;)