In early 2004, preliminary validation was conducted for all SCIAMACHY data products: those generated operationally by ESA's Envisat ground segment and those generated non-operationally at several scientific institutes. Results as presented and discussed during the second Workshop for Atmospheric Chemistry Validation of Envisat (ACVE-2) in May 2004, are summarised here. The main conclusion is that the SCIAMACHY products have improved considerably since the First Envisat Validation Workshop in December 2002, both in number and in quality. All validation results are however still preliminary in the sense that: a) data sets from longer periods with better sampling are necessary to detect seasonally varying errors and other medium-scale effects; b) not all available validation sources were properly used, due to lack of time; and c) comparison results reported here do not always reflect the expected final quality of the products as some of the SCIAMACHY data products are not generated with state-of-the-art processors (due e.g. to delays in the delivery of accurate level-1 data). Various recommendations for product improvement have been collected during the meeting and are summarised here. Note: Due to the short time available for validation before the ACVE-2, several contributions on SCIAMACHY validation in this proceedings have evolved since their presentation at the ACVE-2, so that conclusions for individual products may be different than in the original presentations. This summary paper is however mainly based on the presentations, and not on the papers in these proceedings. Therefore, we refer you to the separate product papers for more up-to-date conclusions on each product.