The sandbox security feature in the next Java Development Kit (JDK) will give programmers even greater control over what downloaded applets can do on their computers and networks, JavaSoft officials said last week.
Li Gong, chief security engineer for Sun Microsystem Inc.'s JavaSoft unit, said the new security model improves on the signed applet model offered in the current JDK 1.1.
The permission-based access control security model in JDK 1.2, due out next April, eliminates the notion of "trusted" remote code such as signed applets, he said.
JavaSoft's signed applet concept was created in response to developers' complaints that the original sandbox security model overly restricted their use of remote code.
The new security model will allow developers to use class loaders that screen local and remote applets to determine what functions they can be allowed to perform on a network.