The provenance of texts entered early into TLS is sometimes unknown, and - correspondingly - the reliability of texts in TLS unfortunately varies a great deal, some of the texts, like that of the Liji and Huainanzi having been inputted by hand, under very difficult conditions, long before such texts were conveniently accessible on the Web. For example, many years ago I spent a large a very large amount of money to have Hanfeizi digitised, and the resulting critical text has many strengths, in fact. But the revised text as it is in TLS now amounts to a new critical edition of the Hanfeizi by Professor Shao Yonghai of Peking University.

For careful critical digital editions of early texts (perhaps over-critical in places), the reader should refer to the CHANT database organised by Professor Ho Che Wah in the Chinese University of Hong Kong. In recent times TLS has been free to use CHANT versions of the pre-Buddhist texts, wherever desirable. For another carefully proofread set of ancient texts there is now always the justly famous old database from the Academia Sinica.

Given the complexities of digitisation the user of TLS is advised always to check quotations from classical Chinese texts against the best printed editions. For example for a text like the Mozi, it is always necessary to refer to the best collated editions in order to ensure that the detail one is commenting on is not in fact textually uncertain. And even for the Lüshichunqiu Wang Liqi's 王利器 new critical four-volume edition 呂氏春秋注疏 (Chengdu: Bashu Chubanshe, 2002) adds a great amount of new insight to the splendid critical editions that precede it.

With certain exceptions, the non-BIG5 variants are represented by the standard character in the computer system so that the epigraphic details are unfocussed for the sake of retrievability, at the cost of exactness of epigraphic transcription.

Christoph Harbsmeier