Subject: Re: [boost] [GSOC] proposal for Trie
From: Hardy (hzithlony_at_[hidden])
Date: 2013-04-26 06:24:38
On 2013/4/26 16:15, Antony Polukhin wrote:
>> Besides, considering that traditional Trie has limited scale of using, I
>> want to design my Trie in 3 layers:
>> The bottom one can provide the user-definition key type, which uses like
>> RawTrie<key_t, value_t>, in which the key can be a sequence of some
>> type(maybe an array of anything can be a key), and every node of Trie can be
>> taken as a root of sub RawTrie, which gives convenience to add Trie into it
>> and even swap a sub Trie.
>> And the second layer will be a specific Trie with the key of chars(when
>> handling lookup and insert, both string and char can be adapted to it).
>> And the feature of sub Trie is provided. The class is like Trie<value_t>.
>> And the third layer will be a Trie without so much freedom of operating as
>> the above two. I will just provide the basic interface as I list above.
>> And the last two layers of Trie will have a version of compressed storage of
> Relations between layers are not clear.
class RawTrie<key_t, value_t>;
provide some interfaces to manipulate sub-tree, and could
A specific RawTrie.
typedef StringTrie<value_t> RawTrie<string, value_t>;
also provide some interfaces to manipulate sub-tree
TrieMap<value_t> - build upon the StringTrie, but can not manipulate the
sub-tree, just provide the interfaces as stl:map and the interfaces I
> Bay the way, student applications can now be submitted to
> The deadline is 3 May.
> Do not forget to submit your proposal and add link to this discussion
> in proposal.
Ok, you mean that I should add this link of our discussion to the
'additional info' of my proposal in GSoC page?
> Best regards,
> Antony Polukhin
> Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk