The third meeting of the Internet Governance Forum in Hyderabad was notable for having introduced a few, mostly long-awaited, innovations that inch the forum further towards the model of an effective multi-stakeholder policy development institution described in the Tunis Agenda. However in each case the limitations or flaws of these innovations have detracted from their potential. Consider the following:

• The publication of a book on the IGF's proceedings. This is one mechanism by which a subset of the IGF's output could, at least informally, filter upwards into other governance institutions. However, being edited by a consultant who is hostile to the idea of the production of recommendations from the IGF, the first edition paints an unbalanced picture of stakeholders' views, particularly in respect of the IGF's own role and mandate.


• The production of recommendations from various workshops and dynamic coalitions, despite the best efforts of some to constrain these bodies into a purely discursive role. For example, the Dynamic Coalition on Open Standards released an Agreement on Procurement in Support of Interoperability and Open Standards, and the moderator of a workshop on The Future of ICANN: After the JPA, What? even put a vote to the floor. But the fact that this innocent and informative poll drew dissent within the MAG is an indictment of that body's commitment to the effectiveness of the IGF as a governance institution.


• The addition of text streaming of the IGF's plenary sessions, to improve access for remote users, and potentially also opening up the possibility of gating this text into other services such as IRC (though, to accomplish this, the non-standard format in which the text is streamed would require a little hacking). But in conjunction with this reform, it has previously been advocated here that in order to be truly interactive the IGF should allow streaming text into, not merely out of the plenary sessions - that is, to interject the live comments of remote users.


• The devotion of afternoon sessions to open dialogue. However, these sessions are unfocused and inevitably unproductive. Until the possibility was scotched by the MAG, driven by the private sector and technical community, it had been proposed that the afternoon plenary sessions would be moderated debates, which could actually have honed the discussions on debated issues towards a tangible result. Whilst even debates are not the appropriate form in which for the IGF's deliberations to take place, they would have been an incremental improvement on what we have been left with.


• Catering for delegates continues to improve. But this is to the detriment of the potential for investment into the IGF's processes. ICANN, the IETF and the RIRs continue to show the IGF up in this respect. Taking ICANN as just one example, it has developed the facilities for automatically-translated multilingual mailing lists, multilingual teleconferencing, and a vibrant public participation Web site (based on the one I co-developed for the IGF in 2006). If half of the extra money spent on food at this year's meeting had been diverted into building an online community for the IGF, participation in the forum could have been greatly broadened and deepened.