AdamS

I could see this being very handy as well. For me some of our clients will only request we do the bare minimum of scope reduction before handing the data to them for further searching, either to keep the billing hours to a minimum or because the data they are searching for is highly specialised in nature and I may not understand it if I see it. In cases like this they will give us a broad scope (date/keywords etc) then we export the data out. I generally manually create the keyword folders and export them out one at a time. The choice to include the same document/email in multiple folders due to multiple keyword hits is a case by case choice depending on my client and the amount of overall data. For small exports I'll generally include the duplicates, for the larger ones I may not. If we could have the option to auto create folders based on tags and export the relevant tags to these folders this would be a great time saver on the backend. If it's possible to have the option to not export the dupes OR export the dupes (tick box maybe), as Intella already has the hash values calculated this shouldn't be beyond possibility. Take that a step further regarding your comment of knowing where in the document the keyword or tag is, if we had the ability to export 2 versions of each document, say one PDF version with the keyword/tags highlighted yellow and then one other 'native' or non highlighted format (Word, PDF, .msg/eml). Which reminds me, where are you guys on having the ability to export emails as .msg? This is something I've been keen to see from the earliest versions as .eml has compatibility issues.

I like the idea of pre-caching in general and it's something I always put forward as an idea for forensic software when I get a chance. I'd love to see this expanded and included as an indexing option for those that are happy to wait the extra hours and also as an stand alone option that can be done at a later time on select objects. For example, a new tab/screen when setting indexing options with some tickboxes for pre-caching for Word, Excel, PDF, Pictures and then a seperate section under the tools menu for select in case pre-caching of tagged objects.

I'm actually in the process of checking with my IT guys if any other traffic is already going out through port 80 that is unrelated to me on that IP address. It may be the best solution all round as port 80 is always allowed through corporate firewalls. Edit : my clever IT guys have set port 443 to route to my server from external connections so now my clients can use port 9999 or 443. Problem solved

I'm curious as to how text analytics could work or be applied. For example if I search for 'pyramid' who or what decides what words are commonly associated with 'pyramid'? Egypt comes to mind and pharaoh, but what if I'm investigating a fraud matter and looking for information around pyramid schemes? While I like two factor auth as a general principal for my banking etc I'm not sure it's really needed for Connect. But if it is implemented at some point it would need to be an opt in rather than default setting I think as it's not something I'd be likely to use. Something I am coming up against on occasion now is clients IT departments being reluctant and in some cases refusing to open port 9999 on their firewall to allow connection to our server. Is it possible to have multiple ports incoming? One client in particular says they only have 4 ports open on the network. I suspect 80, 25, 110 and 22 or 8080 (or what ever their proxy uses).

ahhh yes that's clear. I think one of the Intella guys may need to answer this because I'm not sure how Intella would deal with attachments within a data set. My normal process is to conduct my search, use the include filter to show only emails, then dedupe and export. If the emails themselves are hashing differently due to different content my understanding was that both copies of the email would be included with their relevant attachments, but from what you're saying this may not be what's happening in practice. Edit : although I do seem to recall reading somewhere that this behaviour may be different when we are talking about creating a load file as opposed to simply exporting data.. Edit again: For giggles I just conducted a quick test by finding two emails which are part of a chain both with the same attachment (hash matched), when exporting both these emails the attachments remained in tact for each email. It would appear it may be in issue with regards to creating the load files which I know are far more complicated than a normal export.

It will dedupe everything that is visible in the preview pane. I re-read your point above about 'custodian level dedupe' and I'm curious as to under what circumstances you might need that? Was your client requesting only some of the custodians be deduped and others left alone? If that was the case then you would only have to do two exports, one bulk lot for the deduped and one bulk lot for the not.

You are right in that there is no way to organise custodian sources after they are added, however some sort of linkage would be a great idea. With your 22 cases though you don't have to do 22 different exports, you can do them all in one hit. Under the location facet you can highlight multiple locations by holding the 'cntrl' key as you select (same as windows explorer). Then once the 22 custodians are highlighted, click search then you are good to go. If you are worried about accidentally clicking somewhere and losing the search you can always save that search or just tag them all to create a 'export set', what ever works best for you.

There's 3 wishes in your post With regards to your first point unless I'm misunderstanding what you want here you can already do this. The 'location' facet gives you the ability to search/include/exclude based on custodian already. For your major export with custodian level de-duplication you simply use the location facet, highlight the custodian(s) you are interested in, click search, then de-duplicate, highlight and export. Or have I missed your intention there?

The only problem I can see with a pool approach is that it could cause confusion or problems when users know a specific file exists but they can't find it because it's already been viewed by another user as part of a different search, this would also put an extra burden on the software to track and manage which depending on the complexity of coding that functionality could also impact on overall performance....maybe.... I'm a fan of assigning static batches as for larger cases this is the most efficient way to manage the work load and track the progress for each batch. Likewise with the new data 'merging' seamlessly with old data, while this sounds easier my concern is that in practice it may complicate matters with users not knowing how or where the new data is coming from (assuming a case manager is adding and assigning the data). I'm definitely very keen to see the ability for a case manager to assign and track work flow and I know this is fairly high on the 'to do' list for Connect. I think that the gap between Relativity and Connect is fast closing and IMO Connect is already easier and more effective to use in most cases.

I've had that crop up from time to time too Phil, however only ever duplicated, never the amount you are showing. With only one extra I just right click and removed it from the list to fix my problem, but I can see how that would be time consuming in your case..

Thanks Andrej, although I would respectuflly suggest that the ability to select multiple clusters should possibly be moved onto the high priority list, and in addition when selecting a 'ball' currently there is really no visual clue that the click has been detected. I think if it was made very obvious via highlighting or shading that the 'ball' had been selected this would greatly assist as well.

Two wishes, think I may have mentioned one before.. Email addresses, in Intella there is the ability to select 'all email addresses' then use the dynamic search feature to quickly cut through the clutter. At the moment in Connect we only have the options for 'all senders' or 'all receivers'. Could the 'all email addresses' option be included there along with the dynamic search feature? Also it appears in the latest Connect that we can't select multiple fields to display. By that I mean if I conduct two keyword searches I can't highlght both of the coloured balls to see the combined content, I can only show one at a time.

The hierarchical tags would certainly make it easier, I will hold off further comment until I see it in action and how it appears to the end client, but that sounds like it would do what I'm hoping. With regards to the second part I was thinking that when we select the tags to be hidden it was directly related to the case that was highlighted, but on checking again I think I understand that this tag hiding option is independent of any particular case, is that right? So we can pre-set the tag 'privileged' to be hidden from select users as a global setting, then anytime we use that tag it will be hidden regardless of the case? Which actually makes far more sense when I think about it...

The current backup feature is definitely usable and there are no problems really, it was more the time involved for the backup to happen. I was trying to think of a way that this process could be sped up but then I wasn't considering some other factors like the complexity etc. Certainly something that we can live without, but if some crazy programmer thinks of a way to make it happen it would definitely be a great addition

A couple more wishes for Connect The ability to hide data that is tagged is fantastic, how about adding an option to only hide the tag itself rather than the data it applies to. I'm thinking in terms of removing some visual clutter in cases where I may do the ground work for the client and I have 30 or 40 tags that I have used to create subsets and other various searches that may not necessarily be relevant to the end client. It would be fantastic if at the end of my work I could create a couple of tags that are for the client to review, upload it to Connect and have the ability to hide all my other 'working tags' so they only see the tags I want them to see but can see all the content. OR add an option for permissions where we can specify which tags a particular reviewer can see. By this I mean I could create a new profile caused 'client reviewer' then add a permission that says 'client reviewer' cannot see tags added by user 'adam'. That way when it comes time to add the tags I want to them to see I log in as a different user, add the tags and then they can see only those tags without losing access to any content. Hand in hand with that wish, when we are selecting the tags to hide it would be far simpler if Connect was able to provide a tickbox menu of all the tags currently in use on the selected case rather than manually having to type the tags to hide. You can imagine on a large case that may have hundreds of tags this can become cumbersome particularly when it comes to possible spelling errors which may render the attempt to hide them ineffective.

Pending the official response I think the most important component for a Connect server/host is RAM, more is better. If you are planning on using Connect to undertake the indexing as well (which is coming soon I believe) then CPU and HDD speed/configuration will play an important role as well, but if you intend to index on a separate machine with TEAM then upload to a pure Connect server then RAM is the most important component in my opinion.

Has the search behaviour changed recently? for example if I search for 'resign' this will not be detected within the word 'resignation'. I seem to recall this was the case previously....or did I just imagine it?

I have a large list of client names which I've been given to search across some emails and I've come across a bit of a glitch importing the data, I'll break down the process. I received an excel spread sheet with numerous client names in the following type format J & R Family Trust SMITH, John David Company A Pty Ltd Company B Pty Ltd (The) As there were literally hundreds of company names I used a formula in excel to put quotation marks " " at the beginning and end of each name, I then copied the resulting cells back into a plain text document so I ended up with: "J & R Family Trust" "SMITH, John David" "Company A Pty Ltd" "Company B Pty Ltd (The)" So far so good. However when I import the word list into Intella all the clients with the following format are causing problems "SMITH, John David" These words are not being imported properly, after import they show as "SMITH No closing " and the first and middle names are missing. I can type that exact format manually with no problems but I can't for the life of me figure out why only these ones are importing incorrectly. I should also point out that the quotes don't appear to be causing the issue as when I import the original list sans quotes the same issue is occurring. I assume the problem is being caused by the comma followed by a space, however entering this combination manually is no problems. Is there some part of the import process that ignores text after a comma in plain text documents? As a point of interest removing the comma's fixed the issue so clearly this it what's causing the problems on import.

I'm just curious as to what's happening when we import OCR'd files back into a case. I have 33k files I'm importing now after the OCR process, roughly 25GB of data in this case and I would normally expect the indexing process of a like data set to take about 2-3 hours on my current setup. However the import OCR process I know from experience will take far longer. At this stage the estimate from Intella is 22 hours however I know it will likely take a day or two.

Yes size could quickly become an issue with the structured approach, I didn't think along those lines. However a incremental approach would give much the same and not require much more disk space by comparison, and would have the added advantage of being able to restore to a given day/time based on the incremental schedule.

Everything I've seen suggests that in this case it's simply not in the PST to begin with. It just struck me as unusual, but it may be a by product of extracting PST's from an EDB via exmerge...

I haven't edited the .ini file, everything is set at default. Interestingly on my final attempt that was left running overnight the process completed even though I had the java error visible on the screen at the time.

Server side indexing? Does that mean Connect will be able to index data independent of TEAM?

When I launch the email from Intella into Outlook the name is populated and hyperlinked as described, importing the entire PST into Outlook gives no change. If I select one of the emails in question and hit 'reply' it populates the to field with the name but there is no email address attached to the contact. I'm not thinking this is a limitation of Intella based on what I'm seeing, rather the email address itself doesn't appear to be present at all where the sender is part of the same domain. Could this be something Outlook does by default? Edit: I just had a look at our own internal domain emails to see if this is occurring and it's not, the full email address is included on emails within the domain. We are on 2010 so perhaps this is something that may have changed from 2003 onwards but I'm not able to confirm this. My IT guys haven't any idea why it might be occurring..

Any comments on this Gents? I'm seeing it occur quite regularly and I forsee problems in my future if I can't show the email address completely as "johnsmith" could be anyone without the actual email address somewhere...