Unbalievable waste of money

[railcrew]

Why oh why did you change MW to where it doesn't even do the most basic tasks like detect subject lines like "Viagra", "Please her", "Rolex Watch", all come up as Friend with a big green thumbs up. I "teach" it that those things are spam but the next time they're right back in the green list. I have spamcop and that other one checked but they don't work, I import my friends list from the older version, is shows fine, but disappears after closing and re-opening the program.... So far NOTHING has worked on this other than I check which ones to delete and it does, but I can do that online at the mail client interface. I so want my money back.

[anniebrion]

I find having a couple of simple filters works wonders for the Viagra and watch spams. Also you need about 200 emails to train MWP.

Not sure about the import problem, try building up the friends list from the incoming mails whilst you wait for Firetrust to look into the problem.

IF you tell us your mail server provider Rustic dog might be able to give you a file to get "Origins of spam" working, I have to use a special file because fasthost own antispam cause problems.

[stan_qaz]

Why oh why did you change MW to where it doesn't even do the most basic tasks like detect subject lines like "Viagra", "Please her", "Rolex Watch",

This is a new program and you will have to train it on what is and is not spam. Out of the box it is set to be cautious about marking for deletion so new users don't have mail deleted in error.

all come up as Friend with a big green thumbs up.

If they are coming up as Friend in the status column then you have the address that is on the message in your friends list.

Being on the friends list is not directly related to the thumbs up icon, that is based on the total points scored by the message.

I "teach" it that those things are spam but the next time they're right back in the green list.

If you have the address in your friends list it will take a lot of training to override all the good points that you are awarding the message. If you look at the Bayesian points awarded you should see them trending towards spam as you train.

I have spamcop and that other one checked but they don't work,

Are you not seeing any hits on your messages from the source of spam tools or are they hitting but just not awarding enough points to score the message as spam?

I import my friends list from the older version, is shows fine, but disappears after closing and re-opening the program.... So far NOTHING has worked on this other than I check which ones to delete and it does, but I can do that online at the mail client interface. I so want my money back.

You should have a refund as soon as someone from firetrust sees your post. Sort of sad to see as the new version works a lot better than the old one for folks that take the time to train it and set their points sliders to where they like the results.

[mcullet]

Why oh why did you change MW to where it doesn't even do the most basic tasks like detect subject lines like "Viagra", "Please her", "Rolex Watch", all come up as Friend with a big green thumbs up. I "teach" it that those things are spam but the next time they're right back in the green list. I have spamcop and that other one checked but they don't work, I import my friends list from the older version, is shows fine, but disappears after closing and re-opening the program.... So far NOTHING has worked on this other than I check which ones to delete and it does, but I can do that online at the mail client interface. I so want my money back.

Thanks for the feedback and I'm really sorry that you weren't happy with the new product MWP2010.

Both Stan and Annie have already provided excellent suggestions - won't cut across them.

We all get one chance to make a good first impression - nothing especially clever about that old saying. We need to do better than this and fortunately we can. Whether or not we get a second (third etc) chance to make good impressions ... well, that depends.

Like you, I was hoping to find the latest program (which really is a completely NEW program) worked perfectly with minimal effort on my part. It didn't (as per your example) and I was not happy either. In my case, I probably had unrealistic expectations (based on years of brilliant performance and support by Firetrust) so I have to accept responsibility for my behaviour: this is a tool and not a christmas present (not referring to anyone other than my own perceptions / feelings)

There are plenty (I have one on a thread somewhere) of emails that are so obviously spam / scam that a blind guide dog sitting in the room next door, could tell they were crook.

So why did MWP2010 fail to spot the obvious - and fail is the right word for these emails? Yes, I know we can craft filters and yes, it takes time for the program to learn, and yes there are a slew of settings (ranging from simple to seriously tech level) ... but respectfully to the coders ... the result railcrew describes is one accurate example (of millions of other equally silly results) which cannot BUT give the impression that the new program does not work.

We need to do better than this - really do. The average punter is not really interested in tech issues - they want (whatever) - to work right out of the box. Our efforts to explain what is happening seem really lame IMHO when they are staring at a PENIS ENLARGEMENT spam tagged as GOOD by their brand new purchase. Of course this is going to piss people off - and worst of all, it isn't necessary.

IMHO, while the new program is in learning phase then unless an email address is specifically in a friend list, display the email as NEUTRAL (at least) and ask the user to make a choice (good / bad) which rapidly improves accuracy of spam detection (training). (There are many alternatives to this idea: one example is enough for now.) The reaction to the result railcrew describes is perfectly reasonable (understated): we should pay attention to what happened (he told us) and look at how we can avoid repeating this problem - it is a problem and it goes beyond perceptions.

I'm a geek - and I cannot understand the reasoning behind this self evident silliness. It's like we are deliberately shooting ourselves in the foot here and for not good reason that I can see.

Railcrew (others) - please be patient. All new programs have teething issues. Despite what your eyes tell you, I can assure you that MWP2010 is actually a much better program than the earlier version because of all of the things you cannot see (under the hood stuff). Firetrust (all developers of similar products) must adapt to evolving methods used by the bad guys and this specific program is damn good. It does not help my argument when you are staring at a rubbish result - and I get it. I agree completely with your frustration and understand how you (others) might conclude the new program is utter rubbish.

It's not a rubbish program at all - but we need to address early use issues ASAP otherwise we all look like idiots (as in 'the emperor has no clothes'). Perception is critically important. Then again so too are training the program, filters and so on. These do take some time and effort - please be patient. Andy by all means, please continue to provide feedback.

Railcrew, thanks for pointing out the elephant in the room - and for what it's worth, I'd suggest you hang around, keep coming back: Firetrust do listen and coders are working hard, things are improving and you will see a product that meets your expectations and needs - fairly soon.

[railcrew]

I'll hang in there... but get this... I have (by default) Firetrust in my friends list, I get an email FROM firetrust, it shows as a friend yet MWP classified it as "Undetermined" (Grey)

[mcullet]

I'll hang in there... but get this... I have (by default) Firetrust in my friends list, I get an email FROM firetrust, it shows as a friend yet MWP classified it as "Undetermined" (Grey)

Not sure if you are an Aussie or not ... but your post goes straight to the pool room! It's a reference from a movie called "The Castle" and means, "that was so good it deserves to kept in a sacred place of honour".

Penis enlargement emails evaluate as 'good' - but those from Firetrust are 'neutral'? Yep - cool. I know this is frustrating, but the more we train the program the better (more accurate / faster) it gets. I have a dark hobby in which I go after the scammers / spammers and use one of my catcher accounts (holds > 1,000 nasties) to test performance. Training and filters do work - no argument from me on that. But like you, I believe there is a flaw when obvious spam evaluates as OK when you run the program the first time. I'm working on a post to address the issue because not everyone gets lots of emails (sufficient to reach a sound learning threshold) which means they are relatively unprotected for a while.

Out of curiosity, what version are you running? Mine is a beta 1.0.19 - I can't recall what the latest public release version is ... maybe 1.0.16?

The Firetrust domain is in your friends list - it is in mine, can't recall putting it there but maybe I did ages ago in the previous MWP program?? But unlike you mine evaluates correctly - friend. It's really odd that you got a different result.

Firetrust domain in Friends list

Clip22.png (12.01 KiB) Viewed 13372 times

Emails from Firetrust correctly identified as Friend

Clip23.png (6.25 KiB) Viewed 13372 times

FYI - I did an IP trace on a Firetrust forum email using www.ip-adress.com (yeah - it's the right spelling for the site), and the analysis showed it to come from localhost 127.0.0.1. Don't know if this is what might have triggered the result you saw - perhaps? But if so, the same thing should have happened to me. Really odd.

It might be worth sending your logs to support - just to make sure all is well. (Help / Send Support Logs and follow the pop-up instructions). Maybe there is something wrong with your installation? Dunno - worth the look though.

Thanks for hanging in there - hard as it might seem right now, it really will be worthwhile.

Cheers mate.

[rbrooks]

1. $10 is not a lot of money to "waste."

2. I've been using MW for countless years, so I didn't mind spending $10 to check out the new MW 2010.

3. I did not like having to install .NET 2.51 before installing MW 2010 on my Compaq computer running XP Pro with 2 GB memory (the max possible on my older motherboard).

4. I was surprised and disappointed that MW 2010 didn't import my 30+ spam filters from MW Pro 6.5.4 with boolian ANDs.

5. I did not like screen space that the learning button [good/bad] took.

6. I did not like the non-intuitive UI with the limited HELP that didn't help or the Quick Display that didn't have similar configuration options in the Options menu. Why didn't HELP tell me how to decrease the size of those huge icons that were the default? Why did I have to post a message in the forum to find out where the option was. Why wasn't the "Refer a Friend" option in the main options also? I had to hunt around to figure it out.

7. MS 2010 did not import the SMTP options from my main MW 6.5.4 account. An attempted "Restore" in MW 2010 failed and I had to manually reenter the Recycle Bin Options, Set SMTP server, Advanced options.

7. I only care about functionality, not about 3-D buttons or animated whatevers. MW 2010 was slow to start because of the .NET, seemed to take up too much memory

8. The only feature MW 6.5.4 didn't have was sorting by a header like "Account." After trying it, I decided not to use it.

9. So I uninstalled .NET 2.51 and uninstalled MW 2010.

10. I'm happy to keep my MW licenses for the future possibility of getting a new computer that can handle MW 2010 when it turns into a mature product.

[railcrew]

..." But unlike you mine evaluates correctly - friend. It's really odd that you got a different result."...

It did accknowledge it as "Friend" (in the friend column) but was grey...
I solved the friendlist disappearing problem, I clicked "Clear List" and then re-imported , now it works fine. I added a filter for bad keywords and is learning better now. Thanks

[stan_qaz]

Friend column? There isn't one of them in 2010, do you mean it showed as "Friend" in the status column?

[mcullet]

Hi rbrooks,

(Just saw your email to me - a reply is on it's way soon.)

Thanks HEAPS for this excellent feedback: precise, non-personal, and valuable intelligence. This is stuff we can work with.
No-one can please everyone - you end up with a dog's breakfast.

But fundamental stuff, yep ... we have to address those.

Your comments about 2 gigs of memory being the max available aren't entirely correct. XP is a 32 bit operating system. RAM is a place where programs store 'stuff' - it's very fast. Programs need to remember where they put their 'stuff' and that's where addresses become vitally important. 32 bit operating system can only address (look at) a small amount of RAM dictated by the number of addresses available - can't recall the exact number but it's less than 4 gigs. (Mine has 4 gigs and ignores the extra stuff - happily.) There might be other reasons (warranty related etc ??) that might limit some users from adding extra ram. But in general more RAM does provide real performance improvements - best bangs per buck stuff.

Your points about .NET are well made. My first experiences with it (as a user) came a few years ago when I tried to update a (then) useful commercial utility called Virtual Drive Pro. Unimpressed is a mild description of my reaction. It took the company (and Microsoft) a while to work out coding (from the install process throughout the program itself). That was a few years ago and coders have learned a lot as evidenced by Symantec's product range. BTW, when NIS (Norton Internet Security) and competing products came out (based in whole or in part) on .NET - forums went into meltdown with complaints over performance related issues (speed / resource hogging etc). Over time these got a lot better - NIS works really well on XP and it's only when it's doing a full check that I notice a performance hit - but then, that's what I expect and (for me) not a problem. Accuracy is king.

It's been a while since I installed the .NET runtime on XP but having read your post, I might create a new XP partition and take a long hard look. I recall (perhaps incorrectly) that it's a big download with lots of updates. Most of the updates address security / speed issues - which is good - so long as the end result make a noticeable difference for the end users (you, me ... a slew of people still using XP). I'm HOPING to locate useful (practical) info that improves .NET performance and post it on a thread. It's not a one word answer issue - and you seem to realise that so kudos.

Filter importing - fair cop. The differences between programs (under the hood - how they work) mean that filters from MWP6.x don't automatically translate well to the new program. (Moving stuff from one product to another is called porting.) I'm certain there are threads with 'how-to' instructions to help people do this. IMHO - if we can write 'how-to' instructions to port filters across then we can write a utility to do this which, among other things, might flag those that simply can't work and those that might need a bit of tinkering ... with suggestions. Filters are almost small programs in their own right. I'm lousy at them for the moment but it's on my personal 'things to do list', if for no other reason than I want to understand the thinking (methods?) behind the new program and see if I can create something useful. My point about filters is that it's not always obvious to know the purpose (reasoning / function) of a filter just by looking at it's design - and not all filters are well designed or coded. (No insults to anyone.) I'm acknowledging this is a difficult thing. If it's done poorly (bad filter) then it can have dire consequences.

I've popped a suggestion on the beta forum regarding filters which includes: ability to restore default filters to the default state, save folder for user created filters so we can return to a previous version that worked 'better', and let us share filters more easily between users (file downloads beat keying any day of the week). Filter creation involves experimentation - and IMHO, the current system is fault intolerant. As it, it makes the process needlessly difficult. Others might hold valid alternate views - I'm so new to that area that I'm embarrassed to speak up. But I do - otherwise, what's the point of beta testing anything?

As a beta tester, my approach is, by necessity, different to most users. Among other things, I set out to break the program on purpose. And I have - nothing major but then from a users perspective (which is always my driving guide), a growing list of small annoyances combine to become one BIG problem. No company can fix (rethink in some cases) everything simultaneously - it's not feasible and often makes things far worse (unstable etc). Some things (like cosmetics or minor usability features) might be far more complex than they appear - it all depends on the underlying code.

I'm not Firetrust and I'm still finding my way here - but I can see real value in a public running 'things to do list'. Most people are reasonable. If they can see that an issue has been logged (on the TTDL), status, comments etc - well, it helps improve communications and makes things less personal. Maybe it already exists? I haven't seen it - but then maybe I'm not looking in the right place?

This might help explain where I am coming from here:

Gaming is a multi-billion dollar business. Big money - millions of users world wide. Many of the top companies provide a 'KNOWN ISSUES' list - with info (limited) about when it might be resolved - or not. It does help reduce tension when people see the company (a) knows (b) is working on it (c) has a guide about time-frames for a fix.

We need this type of feedback - it's stuff we can deal with. And if we can't then we will explain why.

I will keep repeating this because it's true. Please be patient. MWO2010 is a new program. The more we train it, the better, faster and more accurate it gets - small time cost - major payoff. We appreciate suggestions but we (me too) have to be patient - some suggestions are mutually exclusive, some might be truly excellent and will require a fair bit of effort to put into reality, some might be low priority.... All companies work to resource constraints: time, talent, money. This is where priorities come into play - and prioritizing is as much an art as a science.

I can tell you (because I see the inter-version change lists) that user feedback is acted upon - it's a big list. I really like what the company does (ethics and products) and it's why I am happy to help volunteer my small contributions towards the project.

Cheers and thanks for hanging in there - please come back from time to time, updates are frequent and we are getting better.

[mcullet]

..." But unlike you mine evaluates correctly - friend. It's really odd that you got a different result."...

It did accknowledge it as "Friend" (in the friend column) but was grey...
I solved the friendlist disappearing problem, I clicked "Clear List" and then re-imported , now it works fine. I added a filter for bad keywords and is learning better now. Thanks

Hey railcrew

I must be more tired than I realised - i just re-read your post (above) and missed a significant issue you described. Status = Friend - evaluates as grey (indeterminate = 'buggered if I know'). (BTW - I too call the status column - friend column, a habit).

If the status was Friend then either the email address or domain exist in your friend list (as in the real friend list). I see email accounts getting hijacked all the time, especially those of people I know or those who have my email address in their contact list. It's a tried and true (and increasing) tactic used by criminals and maybe 'coders for hire'. Lots of hijacked emails appears superficially legit - you see a familiar email address and name and open the door.

ALERT: PLEASE DO NOT TRY TO THIS AT HOME (I do things that are insanely stupid because it's part of my dark hobby - as in, I have tools / methods / mentors and a truckload of security precautions)

Some have nasty payloads - code might be stored on a server in Europe or Korea or China etc. My comment about 'coders for hire' - is real. There are a number of sites where very talented coders bid for work - and complex criminal programs can easily be kitted out by a number of coders - each working on a simple module but none seeing the whole picture. Or some do and dont' care so long as the money is good. You might see a link (all too often it involves shopping - yeah China sites) that are fake - buy this product, go to a shopping cart and give me your credit card details etc. Nope not Amazon or a real name business but (clever / not so clever forgeries). The not so clever sites often have yahoo.com business email addresses. A lot (almost everyone I've tracked came from China).

I hope MWP2010 still examines emails despite the address being in a friend list (don't know). And without seeing the coding of the email in question, maybe 'grey' (indeterminate) was exactly the correct evaluation - and not a glitch.

I apologise for not seeing this sooner - it's important and I missed it. My bad.

[Sundance]

Well, not a complete waste of money - you DO get "First Alert" for 'free'.
I find that everything else is eyewash. I have reverted back to version 6.5.4 as it has everything I need.
But they may keep my $10.00 as I really thing that the previous version is well worth the money.

[mcullet]

Well, not a complete waste of money - you DO get "First Alert" for 'free'.
I find that everything else is eyewash. I have reverted back to version 6.5.4 as it has everything I need.
But they may keep my $10.00 as I really thing that the previous version is well worth the money.

Hi Sundance,

Thanks for your comments - and another I just noticed about BOUNCE issues (http://forum.firetrust.com/viewtopic.php?f=50&t=7333).

Perhaps I'm a heretic - or really tired - or both? But if what you mean is "I'm not thrilled - this is not what I expected and / or It does not work properly" ... then fair cop.

I've identified a few things that just don't make any sense to me - it's like we (Firetrust / users) are shooting ourselves in the foot. Lengthy discourses about 'better' count for precisely dick when I see an email with PENIS ENLARGEMENT (for example) classified as good when the program first cranks up. People aren't silly - and no amount of waffle justifies exposing users to avoidable risk. That's just one of a number of examples ... of avoidable (unnecessary) risky outcomes.

Something is not quite right - perhaps it's in the design - not sure.

I beta test here - but also in a few other places. But unless I have missed something really really obvious - always possible - so far as I can tell, our beta testers are unpaid guinea pigs. What I am doing is not beta testing at all: it's agreeing to expose myself and my system to potentially unsafe software for free, knowing it might result in system harm and agreeing to provide 'feedback' (unspecified / unregulated).

I cannot tell anything I need to know to do the role properly - nothing. My time is preciously valuable to me and while I never claim to have any brilliant skill, I'm no dummy either. Minimum (missing) requirements: absence of obvious formal structures and systems for formal bug reporting, running list of known issues - with report date - status - version - members currently working on the problem - progress notes - suggestions / work-arounds. I have a long list - no point putting it here and I hope my make my point with a few simple examples.

If I report a bug (and there are some 'interesting ones) I can't tell if it's known (unless I read through 100's of posts) when it was first reported or anything of minimal value to me (anyone). I have NO idea if my report goes anywhere or if it's even been noted, tagged for action and so on. For all I know, I'm dropping rocks down a bottomless pit - but we can claim to have beta testers. No - we don't. We have unpaid guinea pigs 'working' without basic systems, structure and direction.

The average punter is not a techo - they deserve and expect to see a product work reasonably well out of the box. As in: if a blind guard dog sitting in a room next door can tell an email is spam but MWP2010 calls it "safe" ... then that's rubbish. Better to get the blind guard dog. I want to help improve MWP and with respect, I might have some modest contributions to make - but I will not waste my time for the sake of appearances and I will not pretend a pile of shite is a rose.

Something is wrong.

Please folks don't mistake my comments as being hostile to Firetrust in any way. I am long term happy user who want to contribute towards making MWP better. We can do a lot better by changing our approaches to beta testing / bug reports etc - really basic project management stuff.

We do need user feedback - and we need to close the feedback loop - info -> action / no action -> sorted? Yes. God. No - place item onto public known issues list which is readily accessible and informative. Users are giving us valuable feedback (which includes their emotional reactions) - we have to respect and honour the effort they took to provide their feedback - and I think we can do much to improve how we do this, and the product.

[therealex]

I've had pretty good luck with MWP detecting spam, and what it does incorrectly is not a big deal to train. What IS completely unacceptable is the lag time when using it. I've updated it to the new patch, and while it's an improvement, it still isn't very good.

I'm using XP Pro SP3 with 4 gigs of memory (yes, I know it can only use 3). I don't mind the $10, since all the previous upgrades were free. But, I just can't see renewing next year. I'll go back to the old MW.

[rusticdog]

I've had pretty good luck with MWP detecting spam, and what it does incorrectly is not a big deal to train. What IS completely unacceptable is the lag time when using it. I've updated it to the new patch, and while it's an improvement, it still isn't very good.

I'm using XP Pro SP3 with 4 gigs of memory (yes, I know it can only use 3). I don't mind the $10, since all the previous upgrades were free. But, I just can't see renewing next year. I'll go back to the old MW.

The lag you mention is the same described here http://forum.firetrust.com/viewtopic.ph ... 52&start=0 ?