Heinrich Cornelius Agrippa (1486-1535) is the most influential writer of Renaissance esoterica, and indeed all of Western occultism. Without doubt, his book de occulta philosophia should be at the top of any required reading list for those interested in Western magic and esoteric traditions. 

Written in three books between the years 1509 and 1510 (he would have been 23 at the time), it was an ambitious attempt to rejuvenate the art of magic which had degenerated during the dark ages. He did this by assembling an intellectual and theoretical foundation from his extensive collection of sources. Agrippa started with a "systematic exposition of ... Ficinian spiritual magic and Trithemian demonic magic (and) ... treatised in practical magic" (I. P. Couliano in Hidden Truths 1987, p. 114). Other major sources used by Agrippa include Liber de mirabilibus mundi of pseudo-Albertus Magnus, Giovanni Pico's Oratio de Dignitate Hominis and Apologia, Johannes Reuchlin's De Verbo Mirifico, Pliny's Historia Naturalis, as well as Picatrix and the Hermetic and Neoplatonic texts. The resulting text circulated widely in manuscript form.]

Over twenty years later Agrippa undertook an extensive expansion and careful revision of the work, which was printed in 1533. Typesetting had scarcely begun before the book was denounced as heretical by the Dominican Inquisitor Conrad Köllin of Ulm. These last minute difficulties account for the inclusion of the lengthy retraction appended to book 3, as well as the absence of the printer's name or location. (Cf. V. Perrone Compagni, Cornelius Agrippa: De occulta philosophia Libri tres, Leiden: E.J. Brill, 1992, p. 11.) 

In his Mysteriorum Libri, John Dee makes frequent mention of Agrippa's book, to the extent that he seems almost to have memorized it. Portions of Agrippa's work are also frequently found appended to magical manuscripts or even liberally merged with the

The English translation appeared in London in 1651. The translator, identified only as "J.F." was probably John French, not J. Freake. (See Ferguson, I, 13 and DNB.) 
In 1801 Agrippa's text, in a slightly abridged form, was shamelessly plagiarized and published as his own work by Frances Barrett (The magus, or Celestial intelligencer, London 1801). This work can still be found in print. The latter was in turn plagiarized and published as his own work by L.W. de Laurence (The Great Book of Magical Art, Hindoo Magic & Indian Occultism, (Chicago, 1915)! He managed the 'Hindoo' part by replacing certain of the Hebrew names with pseudo-Sanskrit fabrications. 
This edition is a transcription of the Gregory Moule edition (Moule: London, 1651.) 
This electronic edition was based on the London 1651 Moule edition found in the British Library:

Title: Three Books of Occult Philosophy ... Translated ... by J. F. [i.e. John French.]    Main heading: AGRIPPA. Henricus Cornelius    Additional headings: F.. J.    Additional headings: FRENCH. John. M.D.    Publication details: pp. 583. R. W. for Gregory Moule: London, 1651. 4o.    Shelfmark: 1471.g.5.    Shelfmark: E.617.    Shelfmark: 232.l.7.(1)  

Graphics were supplied from 1533 Latin edition:   Title:      De occulta philosophia libri tres   Authors:    Agrippa von Nettesheim, Heinrich Cornelius, 1486?-1535.   Published:  [S.l. (Köln/Cologne): s.n., 1533]

I have added text in [] primarily to facilitate searches, but also to include some corrections based on the original Latin (Leiden: E.J. Brill, 1992.) Note the Willis F. Whitehead edition (Chicago, Hahn & Whitehead, 1898) was used in the initial stages of this transcription, but it was found to be less accurate, so I went back and redid the transcription to reflect the earlier edition. His editorial efforts, aside from modernizing spelling, mainly consists of substituting euphemisms for sexual references or deleting them entirely (for examples see chapters 15 and 16). 

The Hebrew lettering in the English edition is full of errors; therefore I have used the Latin Edition (Leiden: E.J. Brill, 1992) to restore these per Agrippa's original intent. Unfortunately, this does not help track errors propagated from the defects in the early English editions.