Alright, I want to address each part of this separately.
Ho-oh
I agree with Maniaclyrisist for the most part. Ho-oh is extremely strong, and with SR as it's only real negative factor it sends a troublesome statement to team builders: use SR if you want any hope of dealing with me. Due to the fact that SR is almost 'required' right now anyway (because of how difficult it is to cover many threats without it), I don't think this alone is a good enough argument to discourage testing. However, what I think will put it over the edge is that fire/flying is a great typing in the current metagame. For instance:
Scizor's Swords Danced Life Orb Bullet Punch vs min/min Ho-oh: 788 Atk vs 216 Def & 353 HP (60 Base Power):153 - 180 (43.34% - 50.99%)
Ho-oh only needs 32HP EVs or 20Def EVs to guarantee survival after SR. If carrying Leftovers, it doesn't need anything. Ho-oh will easily outspeed and OHKO with [fire move], so Bullet Punch is the 'best' option.
Scarf Heatran's Fire Blast vs min/min Ho-oh: 359 Atk vs 344 Def & 353 HP (120 Base Power): 67 - 80 (18.98% - 22.66%)
That doesn't come close to 2HKOing after SR. It needs a Flash Fire boost (and even then Ho-oh could guarantee survival without too much trouble if it wanted).
Infernape's Nasty Plotted Life Orb Fire Blast vs min/min Ho-oh: 614 Atk vs 344 Def & 353 HP (120 Base Power): 149 - 176 (42.21% - 49.86%)
Once again, survives after SR. Fire Blast is literally the best option both those pokemon have against Ho-oh without changing their sets. Maybe they'd start to run HP rock as a result, though. In Infernape's case, Lati@s' existence really hurts it's special sets anyway, so it would likely run SD sets more often than anything.
Those were just a couple calcs I ran out of curiousity. Even assuming SR, Ho-oh with no defensive EVs whatsoever can still serve as a useful check to some of the most used pokemon. No defense EVs whatsoever. Coupled with the rest of the argument (that is: how difficult Ho-oh is going to be to handle once it's in, and the whole 'forces SR' bit), I'm having a hard time believing that a test will be worthwhile.
The only interesting factor I could see here would be that suicide leads might become less popular due to the fact that spinning might become more reasonable as it would allow you to get Ho-oh in easy-style. That would mean people would start relying more on rockers who could actually come back in a few times.
Deoxys-D
Nothing fancy here, I just simply don't see why this is considered to be a suspect. Some of my original serious posting here at Smogon was done because I truly feel that the general playerbase is much more prone to seeing an offensive threat as a problem than a support or defensive one. This is no different, as far as I can tell. It learns a ridiculous amount of support options, and has no trouble setting them up. It's your all-purpose support pokemon, and can run all manner of sets.
I think DX-D best makes this point, though: If not for Tyranitar, would we even be talking about these Psychic types being potentially 'okay' for the standard metagame? Every single time we have one mentioned the main topic of discussion is: 'How does it deal with Tyranitar?' It's even worse than when we say Ho-oh is worthy of consideration due to the presence of SR, because at least SR is a move learned by a good chunk of pokemon. Tyranitar is a single Pokemon.
And no, it's not just Pursuit or a strong STAB Dark move. Weavile and others have a much more difficult time switching in to these pokemon than Tyranitar does, for various reasons.
Mew
I'm absolutely in agreement that Mew isn't a suspect.
Manaphy
This one is really tough. This is one of the few currently Uber pokemon I actually thought deserved a chance originally (without Kyogre, of course!). I still hold true to that, but I would have some concern about the power of Rain Teams in general. It's amazing what a UU rain team can do in standard, but when you add Manaphy to that, things could get crazy. I like the idea of this test, though. Surely, a lot of people would run rain, but at the same time I don't feel it would be centralized toward that due to the many different type of teams available (and weather changing being a good strategy). Also, Manaphy has a big moveslot issue with deciding to carry Rain Dance itself or not.
OHKO/Evasion clauses:
We've beaten these to death and gotten nowhere. I'm going to follow Jump's 'at what point do we start caring about time' and make an appeal here:
Does anyone truly believe that these would help the metagame? Hell, even really change the metagame?
Honestly, I just see these as suspects that we would test with the intention of being both extremely thorough and true to the suspect philosophy. Unlike the top tier 'possibly Uber' threats, these would likely only affect a small part of the game (in what I and others believe to be a purely negative way), and possibly an uncompetitive part anyway.
Species Clause:
I'll admit it, this sounds fun. But that's all it is. Sure, participation wouldn't be as big of an issue (at least at first), but I view this kind of thing as a potential 'fun on the side' metagame rather than anything competitive. There are just too many problems, and such a drastic change in the ruleset means we'd likely have to go through the suspect process for all kinds of things all over again.
Like the OHKO/Evasion clauses, this appeals to the side of me which would like the DPPt metagame to at some point actually be stabilized (rather than the side which is open toward testing everything). In this case, it seems like we'd be testing things for quite a long time.
Soul Dew:
I have my own reservations on just Lati@s themselves, but this is another story. With Soul Dew available, pokemon becomes a 'Do I have to run Tyranitar to deal with Lati@s?' game. The answer seems to be a resounding yes. At the very least, Lati@s in general provoke a very 'necessary' switch in of a physical threat. Even without Soul Dew, there are few 'good' switch ins, and since all of those threaten physical attacks, Reflect is never wasted. This kind of 'forcing the action' is extremely powerful, and when you realize that Lati@s with Soul Dew can also laughably switch into most any special attack with no EV investment at all, it begins to dictate the entire battle on it's own.
We saw a glimpse of this on the Latias suspect test, when double switching out of Latias was ridiculously effective. When playing against such dangerous threats, you just can't afford not to switch in your only hope of dealing with them. Imagine that amplified with Soul Dew, where we would have even more to worry about than just the Dragon-Steel typing war. Mostly because even the steels have trouble taking a Surf/Draco Meteor combo.
No offense here, but I don't really see why we feel this test is necessary. I've noticed that most people seem to agree with that sentiment, and I suggest we don't test Soul Dew at all (especially if Latios is found Uber). If there is truly compelling reasoning to test them, I'd suggest we held off until after we see what happens with Latias in standard play for a while.
Salamence:
Even if IPL is a jerk and doesn't actually explain himself , I'm glad he posted to mention Salamence here. I believe the Latias test opened a lot of people up to the fact that mence may actually fit the idea of Uber 'better' than many of the current Ubers themselves. I myself have a very difficult time explaining why I think certain Ubers are not suspects without in the same breath condemning Salamence. Maybe after thinking it through I can feel more justified in doing so.
A large part of the current standard metagame seems to revolve around how many times Salamence can get in and threaten a switch. There's no doubt that Salamence is the most difficult pokemon to switch into, and you never want to willingly give it a switch in. It's part of the reason we've stopped using the word counter and now say 'check'. When I start thinking of the difficulties of switching into Lati@s, I can only think 'Well, at least it's not Salamence'. It's really (almost) that bad. Mence is usually only shut down by the time you know it's set - which means it's likely killed something already (or will soon). We all know the scenarios, so I don't need to outline it here. Switch the wrong pokemon into the wrong Salamence, and something is dying. Hell, the whole Ice Shard defense pretty much relies on something dying to work in the first place, and at times it seems that mence escapes the Offensive Characteristic only as a result of strong priority moves. The timing of Scizor's Bullet Punch has especially been a huge factor in that.
However, I believe Salamence does fulfill our definition of an Uber in some sense. With SR and priority attacks, it's limited in it's ability to actually sweep your opponent's entire team. On the other hand, it actually provides an inordinate amount of support for your other pokemon due to it's raw power and sheer unpredictability. It either destroys something, weakens multiple pokemon before a true 'check' comes in, or both. Mixmence especially fits the definition, as the only real defense is to 'play around it', which basically means outpredict. I hear this quite a bit too. 'Oh Salamence isn't that bad, just [play around it/use good prediction]' Gengar seemed a bit like this in early DP, but it relied on two 70% accuracy moves to do so (and now that Hypnosis is 60% and Scizor has Bullet Punch, things are even worse for it). It also wasn't much of a physical threat, either.
Realistically though, while some pokemon require a lot more prediction than others, nothing comes close to Salamence. It's not just about the fact it can run CB, Specs, mixed, DD, and other sets either. It's that even if you know it's set, it's still likely to smash you if you predict incorrectly. Again, there's no need for me to make this any more complex than it needs to be - we all know the serious threat Salamence poses, and we don't need to see the calcs for the thousandth time. The real thought here is whether or not we believe Salamence actually fits into the definition that we've decided upon. Are these factors telling enough to consider testing it as a suspect? Does it fit the Offensive Characteristic? Does it fit the Support Characteristic in an unseen way?
To sum up, if it were my decision, I would choose the order of:
Latios
Salamence
Manaphy
-------
Possible Clause Tests after a settled metagame
Latios is obvious, as we're testing that currently. I really think that if there is agreement on Salamence (and I might be reaching on that one), it's more important than Manaphy. I was happy to see that the plans are to test the pokemon before the clauses, as I think we'd all be more willing to sit down and discuss that stuff after we have the tiering worked out. In any case, I feel most of the reasonable suspects have already been tested for the time being, and that we are well on the way to getting toward stage 3 and finding that fabled 'stable metagame'.
Ho-oh
I agree with Maniaclyrisist for the most part. Ho-oh is extremely strong, and with SR as it's only real negative factor it sends a troublesome statement to team builders: use SR if you want any hope of dealing with me. Due to the fact that SR is almost 'required' right now anyway (because of how difficult it is to cover many threats without it), I don't think this alone is a good enough argument to discourage testing. However, what I think will put it over the edge is that fire/flying is a great typing in the current metagame. For instance:
Scizor's Swords Danced Life Orb Bullet Punch vs min/min Ho-oh: 788 Atk vs 216 Def & 353 HP (60 Base Power):153 - 180 (43.34% - 50.99%)
Ho-oh only needs 32HP EVs or 20Def EVs to guarantee survival after SR. If carrying Leftovers, it doesn't need anything. Ho-oh will easily outspeed and OHKO with [fire move], so Bullet Punch is the 'best' option.
Scarf Heatran's Fire Blast vs min/min Ho-oh: 359 Atk vs 344 Def & 353 HP (120 Base Power): 67 - 80 (18.98% - 22.66%)
That doesn't come close to 2HKOing after SR. It needs a Flash Fire boost (and even then Ho-oh could guarantee survival without too much trouble if it wanted).
Infernape's Nasty Plotted Life Orb Fire Blast vs min/min Ho-oh: 614 Atk vs 344 Def & 353 HP (120 Base Power): 149 - 176 (42.21% - 49.86%)
Once again, survives after SR. Fire Blast is literally the best option both those pokemon have against Ho-oh without changing their sets. Maybe they'd start to run HP rock as a result, though. In Infernape's case, Lati@s' existence really hurts it's special sets anyway, so it would likely run SD sets more often than anything.
Those were just a couple calcs I ran out of curiousity. Even assuming SR, Ho-oh with no defensive EVs whatsoever can still serve as a useful check to some of the most used pokemon. No defense EVs whatsoever. Coupled with the rest of the argument (that is: how difficult Ho-oh is going to be to handle once it's in, and the whole 'forces SR' bit), I'm having a hard time believing that a test will be worthwhile.
The only interesting factor I could see here would be that suicide leads might become less popular due to the fact that spinning might become more reasonable as it would allow you to get Ho-oh in easy-style. That would mean people would start relying more on rockers who could actually come back in a few times.
Deoxys-D
Nothing fancy here, I just simply don't see why this is considered to be a suspect. Some of my original serious posting here at Smogon was done because I truly feel that the general playerbase is much more prone to seeing an offensive threat as a problem than a support or defensive one. This is no different, as far as I can tell. It learns a ridiculous amount of support options, and has no trouble setting them up. It's your all-purpose support pokemon, and can run all manner of sets.
I think DX-D best makes this point, though: If not for Tyranitar, would we even be talking about these Psychic types being potentially 'okay' for the standard metagame? Every single time we have one mentioned the main topic of discussion is: 'How does it deal with Tyranitar?' It's even worse than when we say Ho-oh is worthy of consideration due to the presence of SR, because at least SR is a move learned by a good chunk of pokemon. Tyranitar is a single Pokemon.
And no, it's not just Pursuit or a strong STAB Dark move. Weavile and others have a much more difficult time switching in to these pokemon than Tyranitar does, for various reasons.
Mew
I'm absolutely in agreement that Mew isn't a suspect.
Manaphy
This one is really tough. This is one of the few currently Uber pokemon I actually thought deserved a chance originally (without Kyogre, of course!). I still hold true to that, but I would have some concern about the power of Rain Teams in general. It's amazing what a UU rain team can do in standard, but when you add Manaphy to that, things could get crazy. I like the idea of this test, though. Surely, a lot of people would run rain, but at the same time I don't feel it would be centralized toward that due to the many different type of teams available (and weather changing being a good strategy). Also, Manaphy has a big moveslot issue with deciding to carry Rain Dance itself or not.
OHKO/Evasion clauses:
We've beaten these to death and gotten nowhere. I'm going to follow Jump's 'at what point do we start caring about time' and make an appeal here:
Does anyone truly believe that these would help the metagame? Hell, even really change the metagame?
Honestly, I just see these as suspects that we would test with the intention of being both extremely thorough and true to the suspect philosophy. Unlike the top tier 'possibly Uber' threats, these would likely only affect a small part of the game (in what I and others believe to be a purely negative way), and possibly an uncompetitive part anyway.
Species Clause:
I'll admit it, this sounds fun. But that's all it is. Sure, participation wouldn't be as big of an issue (at least at first), but I view this kind of thing as a potential 'fun on the side' metagame rather than anything competitive. There are just too many problems, and such a drastic change in the ruleset means we'd likely have to go through the suspect process for all kinds of things all over again.
Like the OHKO/Evasion clauses, this appeals to the side of me which would like the DPPt metagame to at some point actually be stabilized (rather than the side which is open toward testing everything). In this case, it seems like we'd be testing things for quite a long time.
Soul Dew:
I have my own reservations on just Lati@s themselves, but this is another story. With Soul Dew available, pokemon becomes a 'Do I have to run Tyranitar to deal with Lati@s?' game. The answer seems to be a resounding yes. At the very least, Lati@s in general provoke a very 'necessary' switch in of a physical threat. Even without Soul Dew, there are few 'good' switch ins, and since all of those threaten physical attacks, Reflect is never wasted. This kind of 'forcing the action' is extremely powerful, and when you realize that Lati@s with Soul Dew can also laughably switch into most any special attack with no EV investment at all, it begins to dictate the entire battle on it's own.
We saw a glimpse of this on the Latias suspect test, when double switching out of Latias was ridiculously effective. When playing against such dangerous threats, you just can't afford not to switch in your only hope of dealing with them. Imagine that amplified with Soul Dew, where we would have even more to worry about than just the Dragon-Steel typing war. Mostly because even the steels have trouble taking a Surf/Draco Meteor combo.
No offense here, but I don't really see why we feel this test is necessary. I've noticed that most people seem to agree with that sentiment, and I suggest we don't test Soul Dew at all (especially if Latios is found Uber). If there is truly compelling reasoning to test them, I'd suggest we held off until after we see what happens with Latias in standard play for a while.
Salamence:
Even if IPL is a jerk and doesn't actually explain himself , I'm glad he posted to mention Salamence here. I believe the Latias test opened a lot of people up to the fact that mence may actually fit the idea of Uber 'better' than many of the current Ubers themselves. I myself have a very difficult time explaining why I think certain Ubers are not suspects without in the same breath condemning Salamence. Maybe after thinking it through I can feel more justified in doing so.
A large part of the current standard metagame seems to revolve around how many times Salamence can get in and threaten a switch. There's no doubt that Salamence is the most difficult pokemon to switch into, and you never want to willingly give it a switch in. It's part of the reason we've stopped using the word counter and now say 'check'. When I start thinking of the difficulties of switching into Lati@s, I can only think 'Well, at least it's not Salamence'. It's really (almost) that bad. Mence is usually only shut down by the time you know it's set - which means it's likely killed something already (or will soon). We all know the scenarios, so I don't need to outline it here. Switch the wrong pokemon into the wrong Salamence, and something is dying. Hell, the whole Ice Shard defense pretty much relies on something dying to work in the first place, and at times it seems that mence escapes the Offensive Characteristic only as a result of strong priority moves. The timing of Scizor's Bullet Punch has especially been a huge factor in that.
However, I believe Salamence does fulfill our definition of an Uber in some sense. With SR and priority attacks, it's limited in it's ability to actually sweep your opponent's entire team. On the other hand, it actually provides an inordinate amount of support for your other pokemon due to it's raw power and sheer unpredictability. It either destroys something, weakens multiple pokemon before a true 'check' comes in, or both. Mixmence especially fits the definition, as the only real defense is to 'play around it', which basically means outpredict. I hear this quite a bit too. 'Oh Salamence isn't that bad, just [play around it/use good prediction]' Gengar seemed a bit like this in early DP, but it relied on two 70% accuracy moves to do so (and now that Hypnosis is 60% and Scizor has Bullet Punch, things are even worse for it). It also wasn't much of a physical threat, either.
Realistically though, while some pokemon require a lot more prediction than others, nothing comes close to Salamence. It's not just about the fact it can run CB, Specs, mixed, DD, and other sets either. It's that even if you know it's set, it's still likely to smash you if you predict incorrectly. Again, there's no need for me to make this any more complex than it needs to be - we all know the serious threat Salamence poses, and we don't need to see the calcs for the thousandth time. The real thought here is whether or not we believe Salamence actually fits into the definition that we've decided upon. Are these factors telling enough to consider testing it as a suspect? Does it fit the Offensive Characteristic? Does it fit the Support Characteristic in an unseen way?
To sum up, if it were my decision, I would choose the order of:
Latios
Salamence
Manaphy
-------
Possible Clause Tests after a settled metagame
Latios is obvious, as we're testing that currently. I really think that if there is agreement on Salamence (and I might be reaching on that one), it's more important than Manaphy. I was happy to see that the plans are to test the pokemon before the clauses, as I think we'd all be more willing to sit down and discuss that stuff after we have the tiering worked out. In any case, I feel most of the reasonable suspects have already been tested for the time being, and that we are well on the way to getting toward stage 3 and finding that fabled 'stable metagame'.