I’d like to offer the argument that Wiktionary is the top Latin dictionary available online. I’m sure I’ll meet with the same skeptics who consider Wikipedia an amalgam of amateurs. Let’s fight about it. Here’s my argument.
Wiktionary is superior to two specific alternatives: Whitaker’s Words and Lewis and Short (under Perseus).
Wiktionary is easier to navigate than Perseus or Whitaker. Perseus is notoriously slow, and its entries aren’t incredibly readable. Whitaker is more straightforward, yet demands that I open two windows, which seems unnecessary. There’s no reason a search bar can’t also appear on the results page for further searches.
Wiktionary beats these two because it’s fast and offers a friendly interface.
Like Lewis and Short, it’s pretty good about idiomatic meanings, and it’s OK (just OK) with specifying special case usage and constructions. It’s also often great on etymologies–even pre-Latin etymologies. Admittedly, Lewis and Short are better at both of these features, which is why–I imagine–the link to their entry (under Perseus) is often included at the bottom of each Wiktionary entry. There’s really no need to start with Perseus if you don’t have to.
Whitaker’s Words has a lot of mysterious notation explaining the way in which it parses words (none of which is explained on the site, so far as I can tell). As soon as the site completes its project on era/period frequency, it might be a much better site.
Wiktionary also bests both of them by offering you conjugations and declensions (we’re talking full parses) without asking. They’re just right there on the page. Even better, it’s minimized to begin with, so if you’re a purist and you want to guess the form without the parse, or your just want to scroll down the page a little faster, you won’t encounter opposition.
Wiktionary also shows derived compounds within Latin, and descendant words within other languages. It’s great!
Wiktionary is also the only option to make macrons available everywhere. (If you haven’t noticed, I’m a fan of the macron.)
Of course, if I type ‘lego’ into the search bar, the page includes both legō (to read, etc.) and lēgō (to dispatch), with all the same perks.
Exceptions for using Lewis and Short: (a) where Wiktionary fails to find an idiom or rare usage and (b) where I need to know era/period frequency.
Exception for using Whitaker’s Words: where I have the English and I want the Latin. I haven’t found a better English-to-Latin Thesaurus just yet. Perseus performs this function (in a round about way, and at a snail’s pace). Any recommendations?
Maybe my argument didn’t convince you. Maybe you really like the ‘OLD’ way of doing things. I hate dragging myself to a bookshelf, and I can’t carry that thing to campus. Or, maybe you’re really excited about this:
I’ll stick with Wiktionary.
Latin for Addicts 
Hmm. I’ve never used either Whitacker or Wiktionary and generally use either perseus or my handy-dandy-always-on-my-desk dictionary. Though it is annoying when I’m not at my own computer to look up words. The dictionaries on Perseus are a nightmare to work with. Also, it depends what i’m looking for. For translation, I love this tool: no dictionaries. It allows me to control how much help it gives me, and gives me the chance to read through without help too.
As to the general hate Wiki-things get, I think it’s mostly ridiculous – it’s mistrust generated by those who think it’s trying to be an authoritative source. I’m sure there are few wikipedia articles out there that are down right false, but I’ve never come across any. And seriously, talk about the amount of peer reviewing a wikipedia article gets! It’s certainly possible the article could be biased, but as an educated reader, it’s really your own fault for falling for it – there are lots of biased books and articles out there too! My experience with Wikipedia is as a first stop to familiarize myself with a topic before digging in and exploring the debates and merits of various opinions. Viva la wikithings! 😛 (though as for using wikipedia in academic work, it’s not appropriate not because of the reliability of the article, but because of the tendency for the content to change and the lack of authorship. There is no visible accountability….but I feel like I’m probably preaching to the choir).
I found Nodicitonaries.com back in February, and planned to give it a shot but never got around to it.
https://latinforaddicts.wordpress.com/2012/02/28/nodictionaries-com/
It’s still on my to-do list. Looking back, I see I was worried because it doesn’t feature cases. How do you get around this? Or is not really a problem?
We’re eye-to-eye on Wikipedia. It’s a wonderful resource for the first step, and in a world as data-soaked as our own, sometimes the first step is all we’re after–we don’t have time for anything deeper. Sometimes, we want a simple answer or basic summary, which Wikipedia provides.
Yeah, I do find the lack of a full parse on no dictionaries to be annoying on occasion – mostly when I just need a quick answer. However, when I’m reading something to practice, I find it invaluable – I think it’s a great learning tool when reading for practice, or as a review too for a translation exam etc. (or *shudder* an unseen).
With no dictionaries, I find that I tend to read for sense first. Usually once I have that, the grammar falls into place. It encourages a slightly more natural form of translation than the decoding method we’re usually taught. Speaking of which, the lingua latina books = awesome.