This thread didn't really pass the test of time, as in the 2016 candidates Kramnik started to make the semi-tarrasch the berlin against 1.d4 for awhile in top level chess. (At the moment 1.d4 d5 2c4 e6 3Nf3 Nf6 4Nc3 c5 5c4xd5 c5 xd4  has that "honour" rather than the traditional 5...Nxd5 semi-tarrasch move). 
           Botvinnik's 8Bc4, if white can reach that position there is alternatives for black on the way, still looks quite testing. 10 ...Bf6 is more popular than 10...a6 these days where Smyslov's plan in the last semi-tarrasch in  match against Ribli 1983 Rc1 and exchanging on d5 to play rc7 quite potent.

With so many different move orders, one can expect some transpositional pitfalls. I have always been mainly interested in 1.e4 c5 2.d4 cxd4 3.c3 d5 4.exd5 Qxd5 5.cxd4 Nc6 6.Nf3 e6 7.Nc3 Qd8 8.Bd3 (8.Bc4 here might be even stronger, leading to a superior version of the QGA) Nf6 9.o-o Be7 (this position is treated by Chandler in his books on the Sicilian Alapin) 10.Re1 o-o and according to Chandler both 11.a3 and 11.Bg5 promise an advantage with attacking chances; that lost d4-pawn is often an important weapon: d4-d5!
But in the Semi-Tarrasch Black has more choice: 1.d4 d5 2.c4 e6 3.Nc3 Nf6 4.Nf3 c5 5.exd5 Nxd5 6.e3 cxd4?! 7.exd4 Nc6 8.Bd3 Be7 9.o-o o-o 10.Re1 and Black has other options but Nf6.
Of course such positions are playable for Black. But attacking players should love to play them with White.
I agree, that the 6.e4 is a promising attacking line too. In fact I might prefer this move, as Black can postpone cxd4 after 6.e3 and minimise White's chances.

Regarding Botvinnik-Alekhine1941 our world´s number zero writes in MGP (vol.II, p.123):
"[10...b6?! is] a highly dubios decision. However, as later games showed, White also has a powerful initiative after 10...Nxc3 [...] Apparently the best move is 10...a6".
10...a6 has been played by Karpov, Dreev and other strong players with an equal score. 

These days the Semi-Tarrasch is most often reached via the Caro-Kann or the Symmetrical English (1.c4 c5 2.Nf3 Nf6 3.Nc3 Nc6 4.e3 d5).
Via Caro-Kann it´s easier for White to trick Black into the Bc4 lines than into the Bd3 lines. Because White can play Bc4 before d4 in the lines with 2.c4 but obviously not Bd3 before d4. Playing d4 before Bc4/Bd3 gives Black the opportunity to the nimzo-like Bb4 instead of the more passive Be7.

But even the Be7/Bd3 lines (with cd4 ed4 already played) are playable for Black. It´s not very popular today to analyze passive defending lines for Black so all published material (Aagard, Emms) only shows wonderful victories for White. My own observation is that on 2000-2400 level White makes a phantastic 80% score as most often he is the much better prepared player. But on 2400++ level White achieves no more than his usual 52% as Black knows how to defend. White has the initiative but that´s all!

I really like to defend such positions, White´s Pd4 is simply lost  

The lines with 6.e4 are a quite different topic, obviously many Black players seem to prefer now lines with Nb8-d7. Even Ivanchuk has played this.

tracke  

Two remarks: Black's move order looks imprecise, as White can win a tempo after 6.e3 cxd4?! 7.exd4 Nc6 and now 8.Bd3 at once.
In the game Rasuvajev-Farago, Dubna 1979, Black might be able to improve with Pachman's suggestion 14...Bxg5 15.Bxg5 Qd5.
These are minor quibbles. I agree with Inn2, that every attacking player should rejoice such positions. White just can play natural moves and look for some decisive tactics. I have some experience via the Morra Gambit Declined: 1.e4 c5 2.d4 cxd4 3.c3 d5 4.exd5 Qxd5 5.cxd4 Nc6 6.Nf3 e6 7.Nc3 Qd8 etc. Some books on the c3 Sicilian also have some stuff.

Taimanov in 1980 and Polugajevsky in 1987 think, that Black has reasonable chances for equality if he postpones cxd4 in line 2, eg 6.e3 Nc6 7.Bd3 Be7 8.o-o o-o and White cannot play Bg5 yet.
7.Bc4 Nf6 immediately transposes to the QGA. After 7.Bc4 cxd4 both Taimanov and Polugajevsky give Najdorf-Tal, Belgrade 1970 as the main line. T thinks White has some advantage; P states that chances are equal.
Especially in the IQP positions Markovich is right: White is looking for a piece play based attack against the Black king.
In line one the attack is generally introduced by either d4-d5 or e4-e5.

Anybody any idea where I might get hold of some good information (typical plans and pitfalls) about this opening ? It seems as if there is absolutely no literature available (?!)...