it might not always be the case but must drug that have long half-life have also slow unset, long peak and slow comedown.
No. Plenty of examples to disprove that. Diazepam has a fast onset, relatively short peak effects, but a very long half-life and thus smooth comedown. Oxazepam takes about 2 hours to come on properly, only lasts 3-4 hours, and fades off after that (and has a very short half-life, less than 20 hours).
Look this up, there is 4 level of effects.
1. Anticonvulsant/weaker
2. Anxiolytic/weak
3. Sedative/strong
4. Hypnotic/stronger
http://en.wikipedia.org/wiki/List_of_benzodiazepines
Why did a choice this order ? Simple... Anticonvulsant have to be long acting and make you fonctional, Anxiolytic have to last during the day and make you fonctional, Sedative are similar to anxiolytic but can be strong enough to make you to sleep, While Hypnotic are purely there to make you sleep.
This is wrong on so many levels.
There are not four "levels"; you seem to imply some kind of hierarchy. There are in fact
five classical types of benzodiazepine effect - anticonvulsant, anxiolytic, sedative, hypnotic, muscle relaxant (and amnesic). They are not in any order of strength, but are in fact different effects. Some benzos have a very balanced effects profile, so they are good anticonvulsants, anxiolytics, sedatives, hypnotics and muscle relaxants; diazepam and lorazepam are considered by most to fall into this category. Others have very specific action; triazolam, for example, is mostly hypnotic.
The triple connection you seem to suggest between the different effects being on a linear scale going from weak and long-acting to strong and short-acting is demonstrably false, and doesn't make any sense. Alprazolam is an anxiolytic, yet it has a short duration. Flurazepam is a hypnotic, but it has a very long duration.
You would already notice this if you looked at that Wikipedia link you posted and told me to look at. Check it out - not all anticonvulsants are long-acting, and not all hypnotics are short-acting. The notion that there is a progression like you suggest from weak to strong based on different effects is also manifestly wrong, nor does the source you linked support it in any way.