Word recognition skill is the foundation of the reading process. Word recognition could be accomplished by two major strategies: phonological decoding and sight-word reading, the latter being a marker for proficient reading. There is, however, a controversy regarding the relationship between decoding and sight-word reading, whether the two are independent or the latter is built on the foundations of the former. A related controversy about instructional strategy could be whether to use whole-word method to improve word recognition skills, or to first build decoding skills and then introduce sight words. Five goals were set up to address these issues: (a) developing a criterion that can be used easily by classroom teachers to assess sight-word reading ability, (b) examining this relationship between decoding and sight-word reading, (c) identifying the mechanism that can explain the relationship, (d) examining factors that facilitate sight-word reading, and (e) discussing potential instructional implications of these findings. In order to accomplish these goals, naming time and word-naming accuracy of three groups of subjects (elementary school children, children identified as having reading disability, and college students) were studied by using a variety of verbal materials. The over-all conclusions are that the difference in naming time of letters and words can be used as a metric for assessing sight-word reading skill. Sight-word reading appears to be intimately related to decoding. Sight-word reading is accomplished by parallel processing of constituent letters of words and is influenced also by the semantic nature of words. It is conjectured that sight-word reading instruction is likely to be successful if decoding skills are firmly established first.