Here is my list - it is ranked by potential impact world wide:
1) $100 laptop - not a Linux project per se but if 100 million are distributed this will result in significant impact world wide
2) gnu gcc - the grand daddy of open source projects. Without it we would have a lot less open source software
3) Firefox - xforms, xul, SVG, Javascript 1.6, AJAX
4) Wikipedia - not software but free knowledge for anyone
5) Linux kernel - nuff said
6) Postgresql - will allow governments and small companies world wide to cut their costs and is the back end of Web 2.0. 8.1 + slony has everything any company could need.
7) OpenOffice - mostly a project funded by big IT as a tool to fight off Microsoft - the result is something that everyone can use. Personally I would like abiword + co to lead this category but they simply don't have the momentum.
8) Thunderbird/Sunbird or Chandler – mail and scheduling
8) Twinkle/Kphone Voice is the killer app.
10) JBOSS / Geronimo - projects that if large business embrace will drive down their cost of IT significantly.
Missing/Not developed enought?
1) Streaming Video - we have the codecs but there are not enough solid streaming solutions around
2) Accounting - no J2EE open source accounting package. We have compiere /GNUe - Open Source accounting /ERP. Compiere now runs with postgresql. Accounting and inventory is at the heart of all economic activity. There is also no decent CRM although Chandler could make the basis of a good CRM client.
3) SIP clients - listed but not solid enough
Missed the list?
Gnome is important but is good enough and doesn't run well on small devices. Similarly kde is for traditional pcs. There are dozens of little programs that one might use on a daily basis that also missed the top 10.